当前位置：文档库 › 英文翻译-main

英文翻译-main

A review on computer vision based defect detection and condition assessment of concrete and asphalt civil infrastructure q

Christian Koch a ,?,Kristina Georgieva b ,Varun Kasireddy c ,Burcu Akinci c ,Paul Fieguth d

Dept.of Civil Engineering,University of Nottingham,Nottingham,NG72RD,United Kingdom

Chair of Computing in Engineering,Ruhr-Universit?t Bochum,Universit?tstra?e 150,44801Bochum,Germany c

Dept.of Civil and Environmental Engineering,Carnegie Mellon University,Pittsburgh,PA 15213,United States d

Dept.of Systems Design Engineering,Faculty of Engineering,University of Waterloo,Waterloo,Ontario N2L 3G1,Canada

a r t i c l e i n f o Article history:

Received 22September 2014

Received in revised form 17December 2014Accepted 22January 2015Available online xxxx Keywords:

Computer vision Infrastructure

Condition assessment Defect detection

Infrastructure monitoring

a b s t r a c t

To ensure the safety and the serviceability of civil infrastructure it is essential to visually inspect and assess its physical and functional condition.This review paper presents the current state of practice of assessing the visual condition of vertical and horizontal civil infrastructure;in particular of reinforced concrete bridges,precast concrete tunnels,underground concrete pipes,and asphalt pavements.Since the rate of creation and deployment of computer vision methods for civil engineering applications has been exponentially increasing,the main part of the paper presents a comprehensive synthesis of the state of the art in computer vision based defect detection and condition assessment related to concrete and asphalt civil infrastructure.Finally,the current achievements and limitations of existing methods as well as open research challenges are outlined to assist both the civil engineering and the computer science research community in setting an agenda for future research.

1.Introduction

Manual visual inspection is currently the main form of assess-ing the physical and functional conditions of civil infrastructure at regular intervals in order to ensure the infrastructure still meets its expected service requirements.However,there are still a num-ber of accidents that are related to insuf?cient inspection and con-dition assessment.For example,as a result of the collapse of the I-35W Highway Bridge in Minneapolis (Minnesota,USA)in 200713people died,and 145people were injured [1].In the ?nal accident report the National Transportation Safety Board identi?ed major safety issues including,besides others,the lack of inspection guid-ance for conditions of gusset plate distortion;and inadequate use of technologies for accurately assessing the condition of gusset plates on deck truss bridges.A different,less tragic example is the accident of a freight train in the Rebunhama Tunnel in Japan in 1999that resulted in people losing the trust in the safety and durability of tunnels.According to [2],the failure to detect shear

cracks had resulted in ?ve pieces of concrete blocks,as large as sev-eral tens of centimeters,which had fallen onto the track causing the train to derail.

In order to prevent these kinds of accidents it is essential to con-tinuously inspect and assess the physical and functional condition of civil infrastructure to ensure its safety and serviceability.Typically,condition assessment procedures are performed manually by certi?ed inspectors and/or structural engineers,either at regular intervals (routine inspection)or after disasters (post-dis-aster inspection).This process includes the detection of the defects and damage (cracking,spalling,defective joints,corrosion,pot-holes,etc.)existing on civil infrastructure elements,such as build-ings,bridges,tunnels,pipes and roads,and the defects’magnitude (number,width,length,etc.).The visual inspection and assessment results help agencies to predict future conditions,to support investment planning,and to allocate limited maintenance and repair resources,and thus ensure the civil infrastructure still meets its service requirements.

This review paper starts with the description of the current practices of assessing the visual condition of vertical and horizon-tal civil infrastructure,in particular of reinforced concrete bridges (horizontal:decks,girders,vertical:columns),precast concrete tunnels (horizontal:segmental lining),underground concrete pipes (horizontal)(wastewater infrastructure),and asphalt pave-ments (horizontal).In order to motivate the potential of computer

https://www.wendangku.net/doc/e711256220.html,/10.1016/j.aei.2015.01.008

Handled by W.O.O’Brien

?Corresponding author at:The University of Nottingham,Faculty of Engineering,

Department of Civil Engineering,Room B27Coates Building,University Park,Nottingham NG72RD,UK.Tel.:+44(0)1158468933.

E-mail address:christian.koch@https://www.wendangku.net/doc/e711256220.html, (C.Koch).URL:https://www.wendangku.net/doc/e711256220.html,/civeng (C.Koch).

vision,this part focuses on answering the following questions:(1)what are the common visual defects that cause damage to civil infrastructure;(2)what are the typical manual procedures to detect those defects;(3)what are the limitations of manual defect detection;(4)how are the defects measured;and (5)what tools and metrics are used to assess the condition of each infrastructure element.

Due to the availability of low cost,high quality and easy-to-use visual sensing technologies (e.g.digital cameras),the rate of cre-ation and deployment of computer vision methods for civil engi-neering applications has been exponentially increasing over the last https://www.wendangku.net/doc/e711256220.html,puter vision modules,for example,are becoming an integral component of modern Structural Health Monitoring (SHM)frameworks [3].In this regards,the second and largest part of the paper presents a comprehensive synthesis of the state of the art in computer vision based defect detection and condition assess-ment of civil infrastructure.In this respect,this part explains and tries to categorize several state-of-the-art computer vision methodologies,which are used to automate the process of defect and damage detection.Basically,these methods are built upon common image processing techniques,such as template matching,histogram transforms,background subtraction,?ltering,edge and boundary detection,region growing,texture recognition,and so forth.It is shown,how these techniques have been used,tested and evaluated to identify different defect and damage patterns in remote and close-up images of concrete bridges,precast concrete tunnels,underground concrete pipes and asphalt pavements.

The third part summarizes the current achievements and limitations of computer vision for infrastructure condition assess-ment.Based on that,open research challenges are outlined to assist both the civil engineering and the computer science research com-munity in setting an agenda for future research.2.State of practice in visual condition assessment

This section presents the state of practice in visual condition assessment of reinforced concrete bridges,precast concrete tun-nels,underground concrete pipes and asphalt pavements.2.1.Reinforced concrete bridges

As per US Federal Highway Administration (FHWA)’s recent bridge element inspection manual [4],during a routine inspection of a reinforced concrete (RC)bridge,it is mandatory to identify,

measure (if necessary)and record information related to damage and defects,such as delamination/spall/patched area,exposed rebar,ef?orescence/rust staining,cracking,abrasion/wear,distor-tion,settlement and scouring.While this list of defects comprises the overall list for common RC bridge element categories,such as decks and slabs,railings,superstructure,substructure,culverts and approach ways,not all defects are applicable to all components.Table 1highlights which defects are applicable to which com-ponents and hence need to be checked for each type of component on a bridge.While some of the stated defects are visually detected,some others of them may require physical measurements for accu-rate documentation and assessment.The size of the defect plays an important factor in deciding if it is necessary to go beyond the visual approach.

In addition to the list of defects stated above,FHWA also man-dates that all bearings should be checked during inspection,irre-spective of the material type and functional type of the bridge.Some of the relevant defects for bearings are corrosion,connection problems,excessive movement,misalignment,bulging,splitting and tearing,loss of bearing area,and damage.Furthermore,for seals and joints,inspectors focus on a speci?c set of defects,such as leakage,adhesion loss,seal damage,seal cracking,debris impac-tion,poor condition of adjacent deck,and metal deterioration or damage.While most of these defects can be detected visually,assessing severity of the defects however needs close-up examina-tion and measurements with suitable tools and equipment.

All of the existing defects on a bridge are categorized on a scale of 1–4–each corresponding to the condition state of a particular element (1-Good,2-Fair,3-Poor,and 4-Severe).The condition state is an implicit function of severity and extent of a defect on a com-ponent.Though such categorization of condition states provides uniformity for each component and effects,the actual assessment that results in such categorization can be subjective.Table 2pro-vides some examples of guidelines provided in [4]for categoriza-tion of the condition states of different defects.Please refer to Appendix D2.3in [4]for the complete list of guidelines for all defects.

There are typically three ways to perform manual inspection for concrete bridge elements:visual,physical and advanced.A combi-nation of these methods is required depending on the condition of the bridge member under consideration.During visual inspection,an inspector focuses on surface de?ciencies,such as cracking,spal-ling,rusting,distortion,misalignment of bearings and excessive de?https://www.wendangku.net/doc/e711256220.html,ually,the inspector can visually detect most of the

Table 1

Defects a related to general bridge elements (Grey:Required;White:Not Required)[4]

Del/Spall –Delamination/Spall/Patched area;Exp Rebar –Exposed Rebar;Eff/Rust –Ef?orescence/Rust Staining;Crack –Cracking;Abr/Wr –Abrasion/Wear;Distor –Distortion;Settle –Settlement;Scour –Scouring.

2 C.Koch et al./Advanced Engineering Informatics xxx (2015)

xxx–xxx

relevant defects,provided there is suitable access to the bridge ele-ment.However,visual inspections might not be adequate during the assessment of some speci?c defects.For example,an inspector can identify visually that there is delamination when looking at a patch of concrete,but would not be able to gauge the extent and depth of it accurately by just visual inspection.Visual inspections,without uti-lization of any other inspection techniques,are also known to be subjective which might result in unreliable results [5,6].

In contrast to the visual inspection,efforts during physical inspections are mainly towards quantifying the defects once they are identi?ed visually.For example,to determine delamination areas in a pier or concrete deck,physical methods,namely,ham-mer sounding or chain drag may be used [7].Measurements con-cerning expansion joint openings and bearing positions are also essential during the inspection and evaluation of a bridge.In some cases,advanced inspection methods like those based on strength,sonic,ultrasonic,magnetic,electrical,nuclear,thermography,radar and radiography,are used to detect sub-surface defects or for pre-cise measurements of even surface defects [22].2.2.Precast concrete tunnels

Precast concrete tunnels are one example of civil infrastructure components that are becoming increasingly important when developing modern traf?c concepts worldwide.However,it is com-monly known that numerous tunnels,for example in the US,are more than 50years old and are beginning to show signs of dete-rioration,in particular due to water in?ltration [8].In order to sup-port owners in operating,maintaining,inspecting and evaluating tunnels,the US Federal Highway Administration (FHWA),for example,has provided a Tunnel Operations,Maintenance,Inspec-tion and Evaluation (TOMIE)Manual [8]and a Highway and Rail Transit Tunnel Inspection Manual [9]that promote uniform and consistent guidelines.In addition,Best Practices documents sum-marize the similarities and differences of tunnel inspection proce-dures among different US federal states and transportation agencies [10].

There are different types of tunnel inspections:initial,routine,damage,in-depth and special inspections [8].Routine inspections usually follow an initial inspection at a regular interval of ?ve years for new tunnels and two years for older tunnels,depending on con-dition and age.According to [9],inspections should always be accomplished by a team of inspectors,consisting of registered pro-fessional engineers with expertise in civil/structural,mechanical,

and electrical engineers,as both structural elements and functional systems have to be assessed.However,the focus of this review is on civil and structural condition assessment of precast concrete tunnels.Accessing the various structural elements for up-close visual inspection requires speci?c equipment and https://www.wendangku.net/doc/e711256220.html,mon-ly,dedicated inspection vehicles,such as Aerial bucket trucks and rail-mounted vehicles,equipped with,for example,cameras (used for documentation),chipping hammers (used to sound concrete),crack comparator gauges (used to measure crack widths),and inspection forms (used to document stations,dates,liner types,defect locations and condition codes),are driven through the tun-nel and permit the inspectors to gain an up-close,hands-on view of most of the structural elements.

More recently,integrated and vehicle-mounted scanning sys-tems have entered the market.For example,the Pavemetrics Laser Tunnel Scanning System (LTSS)uses multiple high-speed laser scanners to acquire both 2D images and high-resolution 3D pro-?les of tunnel linings at a speed of 20km/h [11].Once digitized the tunnel data can be viewed and analyzed of?ine by operators using multi-resolution 3D viewing and analysis software that allows for high-precision measurement of virtually any tunnel fea-ture.A different system is the Dibit tunnel scanner that is manually moved through the tunnel [12].It provides an actual comprehen-sive visual and geometrical image of the recorded tunnel surface.The corresponding tunnel scanner software allows easy,quick and versatile data evaluations to visualize the inspected tunnel and manually assess its condition.

According to [9],visual inspection must be made on all exposed surfaces of the structural (concrete)elements (e.g.precast segmen-tal liners,placed concrete,slurry walls),and all noted defects have to be documented for location and measured to determine the scale of severity (Table 3).

Based on the amount,type,size,and location of defects found on the structural element as well as the extent to which the ele-ment retains its original structural capacity,elements are indi-vidually rated using a numerical rating system of 0–9,0being the worst condition (critical,structure is closed and beyond repair)and 9being the best condition (new construction)[9].2.3.Underground concrete pipes

There is a great deal of buried infrastructure in modern cities,most of which appears to be out-of-sight and out-of-mind.Thus,whereas the number of cracks or depths of potholes in asphalt

Table 2

Examples of defects and guidelines for assessment of condition states [4].Defects

Condition states 1234

Good

Fair

Poor

Severe

Delamination/Spall/Patched Area

None

Spall:<1in depth or <6in diameter

Spall:>1in &>6in

diameter;unsound patched area or if signs of distress Situation worse than for Condition State 3and if the inspector deems that it might affect the strength or serviceability of the element

Ef?orescence/Rust staining None

Surface white without

build-up or leaching with-out rust staining Heavy build-up with rust staining

Situation worse than for Condition State 3and if the inspector deems that it might affect the strength or serviceability of the element

Cracking

Width <0.012in or spacing >3ft Width 0.012–0.05in or spacing 1–3ft

Width >0.05in or spacing <1ft

Situation worse than for Condition State 3and if the inspector deems that it might affect the strength or serviceability of the element

Abrasion/Wear No abrasion/wear

Abrasion or wearing has exposed coarse aggregate but the aggregate remains secure in the concrete Coarse aggregate is loose or has popped out of the concrete matrix due to abrasion or wear

Situation worse than for Condition State 3and if the inspector deems that it might affect the strength or serviceability of the element

C.Koch et al./Advanced Engineering Informatics xxx (2015)xxx–xxx

and concrete pavements may very well be the subject of water-cooler conversation,an interest in or an awareness of the state of underground sewage pipes is quite far removed from the percep-tion of most citizens.

However there are two key attributes that motivate attention to underground infrastructure:

1.Being buried,the infrastructure is challenging to inspect.

2.Being buried,the infrastructure is very expensive to ?x or replace.Indeed,the costs associated with sewage infrastructure mod-ernization or replacement are staggering,with dollar ?gures quot-ed in the range of one or more trillion dollars [13].

There is,however,a strong incentive to undertake research and to develop sophisticated methods for underground concrete pipe inspection,due to the huge cost gap between trenchless approach-es and the far more expensive digging up and replacement.The North American Society for Trenchless Technology and corre-sponding No-Dig conferences worldwide demonstrate the wide-spread interest in this strategy,dating back many years [14].

Direct human inspection,which is possible,at least in principle,for above-ground exposed infrastructure such as tunnels and road surfaces,is simply not possible for sewage pipes because of their relatively small size and buried state.Thus there has long been interest [15]in automated approaches,normally a small remote-ly-controlled vehicle with a camera.

A sewage pipe would normally be classi?ed [16]into anticipat-ed structures,

Undamaged pipe.

Pipe joints (connections between pipe segments). Pipe laterals (connections to other pipes).

And some number of unanticipated problem classes:

Cracks.

Mushroom cracks (networks of multiple,intersecting cracks,a precursor to collapse). Holes.

Damaged/eroded laterals or joints. Root intrusion. Pipe collapse.

In common with other forms of infrastructure,the primary challenge to sewage pipe inspection is the tedium of manual examination of many hours of camera data,exacerbated by the sheer physical extent of the infrastructure which,in the case of sewer pipes,exceeds 200,000km in each of the UK,Japan,Ger-many,and the US [17].There are,however,a few attributes unique to sewage pipe inspection:

Lighting is typically poor,since the only light available is that provided by the inspecting vehicle,and any forward-looking

camera sees a well-lit pipe at the sides transitioning to com-pletely dark ahead.

Sewage pipes are subject to extensive staining and background patterning that can appear as very sudden changes in color or shade,giving the appearance of a crack.

Since the focus of this paper is on the computer vision analysis techniques,this following overview of data acquisition is brief,and the reader is referred to substantial review papers [15,17,18,19].Closed circuit television (CCTV)[15,17,20,18,19,21–24]is the most widespread approach to data collection for sewage pipe inspec-tion;nevertheless the sewer infrastructure which has been imaged amounts only to a miniscule fraction of perhaps a few percent [19].Because the most common approach is to have a forward-looking camera looking down the pipe,the CCTV method suffers from drawbacks of geometric distortion,a signi?cant drawback in auto-mated analysis.Sewer scanner and evaluation technology (SSET)[15,16,25,26,19]represents a signi?cant step above CCTV imaging.The pipe is scanned in a circular fashion,such that an image of a ?attened pipe is produced with very few distortions and is uni-formly https://www.wendangku.net/doc/e711256220.html,ser pro?ling [17,27,28,20]is similar to the SSET approach,in that a laser scans the pipe surface circularly,with an offset camera observing the laser spot and allowing the three-dimensional surface geometry of the pipe to be constructed via triangulation.

There are a few further strategies,albeit less common,for sewer pipe inspection.A SONAR approach [15,28,19]has been proposed for water-?lled pipes,where most visual approaches will fail,par-ticularly if the water is not clear.Ultrasound methods [29–31,17],widely used to assess cracks in above-ground pipes,have been pro-posed to allow an assessment of crack depth,which is dif?cult to infer from visual images.Infrared Thermography [15,17,19]relies on the fact that holes,cracks,or water intrusion may affect the thermal behavior of the pipe and therefore be revealed as a ther-mal signature.Finally,ground penetrating radar [15,17]allows the buried pipe to be studied from the surface,without the clutter and challenges of driving robots in buried pipes,but at a very sig-ni?cant reduction in resolution and contrast.

Because of rather substantial cost associated with data acquisi-tion of sewer pipes,there is signi?cant interest in maximizing the use of data.Prediction methods [32–35]develop statistical,neural,or expert system deterioration models to predict pipe state,over time,on the basis of earlier observations.2.4.Asphalt pavements

As reported by the American Society of Civil Engineers (ASCE),pavement defects,also known as pavement distress,cost US motorists $67billion a year for repairs [36].Therefore,road surface should be evaluated and defects should be detected timely to ensure traf?c safety.Condition assessment of asphalt pavements is essential to road maintenance.

Table 3

Common civil/structural defects of concrete tunnels and respective severity scales according to [9].Defect type/Severity Minor Moderate

Severe

Scaling <6mm deep 6–25mm deep

>25mm deep

Cracking

<0.80mm

0.80–3.20mm,or <0.10mm (pre-stressed member)

>3.20mm,or >0.10mm (pre-stressed member)

Spalling/Joint Spall <12mm deep or 75–150mm in diameter 12–25mm deep or $150mm in diameter >25mm deep or >150mm in diameter

Pop-Outs (holes)<10mm in diameter 10–50mm in diameter

50–75mm in diameter (>75mm are spalls)Leakage

Wet surface,no drops

Active ?ow at volume <30drips per minute

Active ?ow at volume >30drips per minute

4 C.Koch et al./Advanced Engineering Informatics xxx (2015)

xxx–xxx

There exist several techniques to detect distress in asphalt pavements.These techniques differ in the pavement data which is being collected and in the way this data is processed.Sensor-based techniques utilize devices to measure parameters of the pavement surface.Visual-based techniques make use of observa-tions of the pavement surface to identify anomalies that indicate distress.Depending on the way of processing data,techniques are classi?ed as purely manual,semi-automated or automated [37].Manual processing is entirely performed by experts,while semi-automated and automated techniques require little or no human intervention.

Visual-based techniques consist in manually inspecting the road surface or employing digital images and computing devices to assess the pavement condition.In case of manual inspection,trained personnel walks over the road shoulder and rates the pave-ment condition according to distress identi?cation manuals.The disadvantage of this technique is that it is subjective despite the use of manuals and it depends on the experience of the personnel.Also,the personnel are exposed to traf?c and weather,which makes the inspection procedure hazardous.Another issue related to the manual inspection of the road service is the time required to perform it.

To speed up the assessment process,pavement images are ana-lyzed instead of walking on the roads.Pavement images are obtained using downward-looking video cameras mounted on sophisticated vehicles.When the images and data are analyzed by human experts,the process of assessing the pavement condition is semi-automated.However,the rating of the pavement still depends on the experience of the analyzer and the subjectivity issue remains.

Most distress detection techniques,regardless of whether they are manual,semi-automated or automated,depend on the pave-ment distress type.Pavement distress varies in its form and https://www.wendangku.net/doc/e711256220.html,monly,distress is characterized as alligator cracking,bleeding,block cracking,depression,longitudinal or transverse cracking,patches,potholes,rutting,raveling and more.The U.S.Army Corps of Engineers,for example,distinguishes between 19types of dis-tress [38].

Distress types and measurements are de?ned in visual pave-ment distress identi?cation manuals.Some of these measurements and indices vary between different countries,and federal states.Table 4presents examples of defect assessment measurements and condition indices de?ned in such manuals [39–43].As can be seen,severity and extent are present in most of the manuals.The common procedure to obtain the extent value is to count the occurrences of the different severity levels for each type of distress for the whole segment and convert the amount of distress into dis-tress percentage.

Condition assessment indices are calculated based on the dis-tress measurements.Several pavement condition assessment indices have been developed and the procedures of their calcula-tion are described in visual distress identi?cation manuals.For instance,the pavement condition index (PCI)is widely used.The pavement condition index is a statistical measure of the pavement condition developed by the US Army Corps of Engineers [38].It is a numerical value that ranges from 0to 100,where 0indicates the worst possible condition and 100represents the best possible con-dition.A verbal description of the pavement condition can be

de?ned depending on the PCI value.This description is referred to as pavement condition rating (PCR).PCR classi?es the pavement condition as failed,serious,very poor,poor,fair,satisfactory or good.

https://www.wendangku.net/doc/e711256220.html,puter vision methods for defect detection and assessment

This section presents a comprehensive synthesis of the state of the art in computer vision based defect detection and assessment of civil infrastructure.In this respect,this part explains and tries to categorize several state-of-the-art computer vision methodolo-gies,which are used to automate the process of defect and damage detection as well as assessment.Fig.1illustrates the general com-puter vision pipeline starting from low-level processing up to high-level processing (Fig.1,top).Correspondingly,the bottom part of Fig.1categorizes speci?c methods for the detection,classi?cation and assessment of defects on civil infrastructure into pre-process-ing methods,feature-based methods,model-based methods,pat-tern-based methods,and 3D reconstruction.These methods,however,cannot be considered fully separately.Rather they build on top of each other.For example,extracted features are learned to support the classi?cation process in pattern-based methods.Subsequently,it is shown,how these methodologies have been used,tested and evaluated to identify different defect and damage patterns in remote and close-up images of concrete bridges,pre-cast concrete tunnels,underground concrete pipes and asphalt pavements.

3.1.Reinforced concrete bridges

Much of the research in defect detection and assessment using computer vision methods for RC bridges have largely focused on cracks,and to some extent on spalling/delamination and rusting.Many of these research studies targeted and contributed success-fully to the automation of detection and measurement of defects.More studies need to be done to improve the methods used for automatic assessment as they are currently based on several assumptions.

In addition to cracks,there are also other defects that are essen-tial to be detected and assessed in relation to a RC bridge.Being able to detect,assess and document all defects as independent entities is paramount to provide a comprehensive approach for bridge inspection.

Currently,some of the other categories of defects are being inherently detected or assessed as part of other major dominating defects present at the using computer vision methods.For exam-ple,some methods detect abrasion as part of the crack [44].In other cases,such as distortion and misalignment of bearings,no automated method exists for detecting and assessing them.This clearly indicates that more research needs to be done in the direc-tion of automating the detection and assessment of various defects.To be able to perform automatic assessment and condition rating assignment,as a ?rst step,it is necessary to identify the relevant defect parameters to accurately and comprehensively represent the defect information.

Below we will present the synthesis of the research done so far in the computer vision domain for various types of defects.

Table 4

Examples of pavement defect assessment measurements and condition indices.

Ohio [39]

British Columbia [40]Washington [41]

South Africa [42]Germany [43]

Measurement Severity,extent

Severity,density

Severity,extent

Degree,extent

Extent

Index

Pavement condition rating

Pavement distress index

Pavement condition rating

Visual condition index

Substance value (surface)

C.Koch et al./Advanced Engineering Informatics xxx (2015)xxx–xxx

3.1.1.Cracking

Previously,Jahanshahi et al.[45]reviewed automatic defect detection approaches.Very recently,Rose et al.[46]reviewed existing crack detection and assessment algorithms for concrete bridges and classi?ed them broadly as edge detection,segmenta-tion and percolation,machine learning methods,morphology operations,ground and aerial robot photography,template match-ing,and other techniques.Building on this categorization,we reviewed and discussed some of the existing algorithms below.

Abdel-Qader et al.[47]compared various edge detection algo-rithms and found the Haar Wavelet method to be the most reliable among them,for the purpose of crack detection.However,the per-formance of edge detection algorithms on noisy image data is questionable,and same is the case with morphological operation based methods[48].Yamaguchi et al.[49]used scalable local per-colation-based image processing techniques and they proved to be ef?cient and accurate even for large surface images[50].Abdel-Qader et al.[51]used a Principle Component Analysis based algo-rithm to detect cracks on a bridge surface.In this case,the accuracy of results varied with camera pose and distance from where images are taken.Prasanna et al.[52]developed a histogram-based classi?cation algorithm and used it along with Support Vector Machines to detect cracks on a concrete deck surface.The results of this algorithm on real bridge data highlighted the need for improving the accuracy.Nevertheless,training data from various locations on the bridge could be used to build the classi?er and testing could be done on data from a different location of similar structural composition.Similarly,Lattanzi and Miller[53]devel-oped an automatic clustering method for segmentation based on Canny and K-Means to achieve greater accuracy of crack detection under various environmental conditions at a greater https://www.wendangku.net/doc/e711256220.html,t-tanzi and Miller’s work is signi?cant,especially if training data comprises images from different locations because it is important to offset the environment variability associated with variable light-ing and shading conditions at different locations on the bridge, which is often the case with real world bridges.Some researchers also combined image-based3D scene constructions with other techniques,in order to obtain depth perception that a2D image lacks,to support automatic crack detection[54,55].

While the above algorithms demonstrated capabilities to detect cracks,it is also important in a bridge inspection to understand the crack properties such as location,width,length and orientation, because condition ratings for bridge elements are assigned based on such properties.As outputs of the process of extracting proper-ties from images are quantities,it is imperative that images are mapped to the global coordinate system.This requirement stems from the likelihood that images are collected on?eld with varying con?gurations,i.e.resolutions,positions,orientations,etc.,over different inspections,which is primarily due to dif?culty in repli-cating the same image capture con?guration as well as a result of rapid advances in camera technologies over relatively shorter time periods.Towards normalizing different images to true world scale,different researchers used techniques such as3D pose esti-mation,multiple image stitching or by making measurements rela-tive to the host structural element.In relation to that,some data acquisition systems used by researchers also had3D pose control feature.These systems likely comprised surface-based(ground-based,water-based,bridge surface crawler)or aerial robots,which can either,have pre-con?gured settings or can log accurate image capture con?guration dynamically.

Targeting to achieve the goal of going beyond mere crack detec-tion,Yu et al.[56]developed a graph-based search method to extract crack properties for further assessment and used a ground-based robot for collecting images;however,this method

Pre-processing Segmen-

tation

Feature

extraction

Object

recognition

Structural

analysis

restoration, contrast enhancement, noise reduction thresholding,

edge detection,

region growing,

clustering

features

related to

color, texture,

shape, motion

classification

of materials,

objects,

shapes,

locations spatio-temporal scene understan-ding,

Median

filtering Morphologic operations Histogram equalization Image stitching Shadow removal

Haar-

wavelets

Hough

transform

Canny Histogram of Oriented

Gradients(HoG)

Background substraction

Laplacian of Gaussian

(LoG)

Wavelet

transform Garbor filtering

Pre-processing Feature-based

methods

Model-based

methods

Pattern-based

methods

3D Recon-

struction

Percolation-based

models

Line-tracing

algorithms

Generalized Hough

transform

Graph-based

Multi-temporal

methods

Principle Com-

ponent Analysis

Nearest

neighbor

Support vector

machines

Neural

networks

Stereo

reconstr.

Multi-view

reconstruction

Image

registration

Optical

flow

Image stitching

computer vision methods(top)and speci?c methods to defect detection,classi?cation and assessment of civil infrastructure.

6 C.Koch et al./Advanced Engineering Informatics xxx(2015)xxx–xxx

needed manual input of start and end points of crack [50].Later,Oh et al.[57]demonstrated a technique implementing automatic two-step:crack detection and crack tracing algorithm to be able to detect as well as identify crack properties,such as width and length,and tested the developed algorithm on a real bridge.They collected images with a ground-based robotic system that had con-trolled pan and tilt mechanisms,and used median ?lter for smoothening in the pre-processing stage,then isolated the candi-date crack points and applied morphological operations such as dilation and thinning to maintain crack segment connectivity.As part of their study,they compared their results with Fujita,Sobel and Canny’s method.The performance of the algorithm proposed by Oh et al.[57]matched the other three methods in terms of eliminating shaded regions and detecting major cracks,while out-performing them in the case of thinner cracks.

Other researchers targeted developing crack maps.Jahanshahi et al.[58]proposed a crack detection system to extract a complete crack map using 3D scene reconstruction,morphological opera-tions and machine learning classi?ers,and followed it up with a robust photogrammetry-based approach to compensate for cam-era perspective errors [59].In another recent case,Zhu et al.[60]proposed a novel method involving thinning of the crack maps and subsequent measurement of each crack skeleton point to the crack boundary to automatically extract necessary crack para-meters [50].More recently,Lim et al.[61]proposed a Laplacian of Gaussian (LoG)based algorithm to perform crack detection and mapping on an RC bridge deck,and uses a mobile robotic sys-tem that can traverse a deck surface to capture images.The robot stores the spatial locations of image capture and uses robot coordi-nate system to transform from image coordinate system to global coordinate system.

The results presented in most of these cases were based on application of their methods on bridge deck surface,or in some cases image data of the beams and columns were considered.Gen-erally speaking,most of the images used in these studies were images from simple ?at and curved surfaces.However,the joints,seals,bearings and other connections present more complex geo-metry,often comprise of many sub-components and generally have varying material composition.Thus,these conditions render it hard to distinguish cracks from true edges.Also,as bridge inspec-tors commonly look out for connection related defects,algorithms should be tested on images from these components.

3.1.2.Delamination/Spalling

Only recently,there have been developments in the detection and assessment of spalling on concrete surface and these works seem to have drawn inspiration from rusting detection and assess-ment [50].German et al.[62]considered a combination of segmen-tation,template matching and morphological pre-processing,both for spall detection and assessment on concrete columns.They identi?ed length of spalled region along longitudinal direction and distance between exposed reinforcement bars in the trans-verse direction and developed an approach for assessing the cumu-lative severity of the spalling based on different enumeration levels –(i)spalling of concrete cover without exposing reinforcement,(ii)spalling exposing longitudinal reinforcement and that of core con-crete.The results obtained for the test images indicated spall detection with a precision of 81.1%and a recall of 80.2%for a set of 70images.However,they indicated that more work is needed to achieve more detailed categorization of spall property result,with particular focus on spalling that exposes transverse reinforcement.

Adhikhari et al.[63]presented a novel approach based on orthogonal transformation,using shape preserving algorithms such as af?ne and projective transformation,to overcome perspec-tive and parallax errors of a camera during data collection that can

result in inaccurate defect quanti?cation.They could determine if spalling had occurred,and if spalling was present,they could retrieve spall properties automatically.Their research also used Bridge Condition Index (BCI)after quantifying the defects to map them to condition ratings.While they could achieve reasonably accurate results (85%accuracy)for automatic procedures,their algorithm could not completely address automatic identi?cation and assessment in situations where multiple defects (e.g.spall and crack)interact at the same spatial location.

Though work on spalling detection and assessment started only recently,the progress so far is very promising.The algorithms have been tested with images from decks and columns.Like in the case of cracks,even spalling needs to be checked for at concrete joints.Therefore,including images from those locations will be valuable for better detection and assessment performance of the algorithms.

3.1.3.Other damage scenarios

Zaurin et al.[3]used video imagery and bridge responses col-lected by strain gauges and fused them together to detect loss of connectivity between different composite sections,and change in boundary conditions.In the process,unit in?uence line of the bridge is extracted and statistical outlier detection is done to dif-ferentiate damage state from the baseline state.This method was tested using a four span experimental bridge belonging to Univer-sity of Central Florida.Adhikari et al.[64]presented an change detection approach based on fourier transformation of the images,which could useful for detecting subtle defects such as periodic and sudden settlement of substructure.The review of the paper also suggests no proper basis for thresholding,and the results vary depending on the chosen threshold limit chosen.However,this method is a signi?cant improvement over traditional change detection approach using the image difference,and can be used to quickly do a temporal comparison of different images.Uhl et al.[65]developed a method to detect de?ection in structural members by applying homography mapping.Speci?cally,they implemented an automatic shape ?lter and a corner detector to calculate the de?ection using homography mapping between the two views.They implemented this algorithm on an experimental set up in a lab,and also on a real bridge,and veri?ed their results with the de?ection calculated using a laser scanner.The results seem to be very accurate with the average difference between both the measurements being less than 0.5%.Though the de?ection is being calculated accurately,it did not address the problem of dam-age localization and assessment.Kohut et al.[66]extended Uhl et al.’s work [65]to include damage localization and assessment using a wavelet transforms based analysis method to do irregular-ity detection.

Various algorithms,related to detection and assessment of cracking,spalling and some damage scenarios in RC bridges,have been discussed above,and our focus was on the progress of the computer vision research in terms of automation in detection and assessment of these defects.

3.2.Precast concrete tunnels

In contrast to concrete bridge inspection,the image and video data acquired inside a tunnel is much different in terms of arti?cial lighting and camera distance.From that perspective,it is interest-ing to review the current state-of-the-art computer vision algo-rithms for defect detection in tunnel image data.According to Chaiyasarno [67],automated tunnel inspection systems that cover both defect detection and condition assessment can be grouped into the following themes:detection,visualization and interpretation.

C.Koch et al./Advanced Engineering Informatics xxx (2015)xxx–xxx

3.2.1.Defect detection

In analogy to concrete bridges,the most sought after defects are cracks as they are the primary indicator of deterioration patterns,which are due to other severe causes that need to be further ana-lyzed [68].Yu et al.[56]also highlight that cracks are of particular concern as they most signi?cantly affect the state of the concrete within a tunneling environment.

Computer vision methods for crack detection generally involve a pre-processing step and a crack identi?cation step.First,in the pre-processing step image processing techniques are applied to extract potential crack features,such as edges (threshold-based approaches).Second,the identi?cation step usually applies crack modeling (model-based approaches)and/or pattern recognition techniques (pattern-based approaches)in order to classify if the extracted features belong to crack regions.Next to methods described in the previous section,mentionable contributions that are applicable to crack detection during tunnel inspection are the described below.

3.2.1.1.Threshold-based approaches.Miyamoto et al.[69]calculate the difference in intensity between each pixel and the average intensity of each row in an image.A pixel that differs considerably from the average is said to be a crack pixel.Fujita et al.[70]use a line ?lter based on the Hessian matrix to emphasize line structures associated with cracks before they apply thresholding to separate cracks from background.

The major drawback of threshold-based approaches is the ques-tion on how to choose a suitable threshold for extracting crack fea-tures.The described algorithms select a threshold based on prior knowledge.However,such methods can hardly be generalized and may be inapplicable to the imaging conditions found in real tunnel images.Moreover,they are prone to inaccuracy caused by shadows as the intensities of shadow pixels tend to have a similar brightness compared to crack pixels.

3.2.1.2.Model-based https://www.wendangku.net/doc/e711256220.html,ai [71]developed a crack detec-tion system based on the deformation of tunnel walls.Under this method,the model of a crack is characterized by eight quantities,such as area and Feret’s occupancy rate.Subsequently,a ?lter is used to remove noise.Yamaguchi et al.[49]modeled cracks based on the concept of percolation,which is a physical model describing the phenomenon of liquid permeation.The algorithm starts by ini-tializing a seed region and then the neighboring regions are labeled as crack regions based on the percolation process.Paar et al.[72]present a crack detection algorithm based on the line tracing algo-rithm that assumes a crack is a series of short straight lines con-nected together.Again,the algorithm starts from a seed point followed by searches for a line within the neighboring regions.Yu et al.[56]proposed a crack detection method in conjunction with a mobile robot system for automated inspection of concrete cracks in tunnels.Their method calculates the length,thickness and orientation of concrete cracks through a graph search;howev-er,it requires the crack’s start and end point to be manually pro-vided.Moreover,the robot is required to maintain a constant distance from the tunnel wall in order to achieve accurate mea-surements of the damage properties.This system claims to have an overall detection accuracy rate of 75–85%and a measurement error of recognized cracks of less than 10%.

According to [67],model-based methods for crack detection strongly rely on user input to initialize the seed pixels.Conse-quently,hairline cracks may not be detection because users may be unable to identify the seed pixels.Due to reliance on the user input,these methods may not be scalable.

3.2.1.3.Pattern-based approaches.Liu et al.[73]apply a Support Vector Machine (SVM)classi?er to determine if crack features

appear in an image patch.Potential crack features are pre-de?ned based on intensity.Abdelqader et al.[51]use a Principal Compo-nent Principles (PCA)algorithm that reduces the dimensions of fea-ture vectors based on eigenvalues,and extracts cracks from concrete images.The images are ?rst pre-processed by line ?lters in three directions:vertical,horizontal and oblique;then further processed by the PCA algorithm and classi?ed based on the nearest neighbor algorithm.

Methods based on pattern recognition considerably rely on training data in order to set up robust classi?ers.Training and validation data are usually performed by manual labeling (super-vised learning),which is a labor-intensive and error-prone procedure.

3.2.2.Visualization

The main goal of visualization is to visually organize large image and video data sets to enhance inspection.Image stitching or image mosaicing is a common method to combine and visualize a collection of images.In the domain of tunnel inspection,Chaiya-sarn et al.[74]present a system that constructs a mosaic image of the tunnel surface with little distortion.Their system obtains a sparse 3D model of the tunnel by multi-view reconstruction [75].Then,the Support Vector Machine (SVM)classi?er is applied in order to separate image features lying on the cylindrical surface from those of the non-surface.The reconstructed 3D points are reprojected into images for accurate cylindrical surface estimation.Jahanshahi et al.[76]create stitched images of structural systems from a specialized camera that can tilt and pan.The method detects missing parts,such as bolts,when comparing images taken at different times for the purpose of structural health monitoring (SHM).The method applies a machine-vision algorithm to perform image registration to rectify images so that they are in the same coordinate frame.

In general,image stitching provides a feasible way of increasing the ?eld of view that cannot be achieved by a single image.Conse-quently,a wide-angle or stitched image may improve defect detec-tion results,in particular in case of hairline cracks,since the stitched image provides a higher resolution of defects,e.g.cracks.3.2.3.Change monitoring

Apart from detecting cracks,classifying crack patterns and asso-ciated sizes,it is essential to observe if cracks in tunnel liners have changed over time and how quickly they do so.This kind of infor-mation helps determine the deterioration rate of the structural tunnel components [67].

Lim et al.[77]propose a system for change monitoring of cracks from multi-temporal images.Their system is based on a 2D projec-tive transformation that can accurately determine the crack size,which is then monitored in consecutive images as the crack propa-gates.Although this system that can cope with images taken from different viewpoints,it requires explicit user input for the control points,which makes the system unscalable for a large number of images.Chen and Hutchinson [78]propose a framework for con-crete surface crack monitoring and quanti?cation.Their method is based on optical ?ow in order to track the movement of cracks.However,current solutions related to monitoring cracks or anoma-lies rely greatly on some degree of user input [67].3.3.Underground concrete pipes

Deplorably,on the basis of a search of sewage pipe inspection methods currently offered by North American contractors,most buried pipe inspection continues to be manual and CCTV based,implying a slow inspection process subject to operator fatigue and boredom.Although this limitation is frustrating,it strongly motivates continued research work on machine intelligence and

8 C.Koch et al./Advanced Engineering Informatics xxx (2015)

xxx–xxx

computer vision in this application,and is the driving motivation for this section.There have been signi?cant with computer vision contributions to pipe inspection,in whole integrated systems such as PIRAT [28,18],KARO [18],and AIMP [18,79],and the mapping the underworld (MTU)project [19].

The computer vision analysis of underground concrete sewer pipes has much in common with other forms of infrastructure.In particular,all of the parallel sections in this paper discuss aspects of crack detection,hole detection,and the classi?cation of cracks into different forms or degrees of severity:multiple cracks,net-worked cracks etc.The forms of concrete deterioration in different parts of infrastructure do,after all,share a great deal in common.As discussed in Section 2.3and in review articles [15,17–19],an unusually wide variety of possible imaging modalities has been developed for buried pipe inspection.In terms of the role of com-puter vision,we will focus our discussion on the most widespread methods,which have seen the most attention in the literature,namely the CCTV,SSET,and laser pro?ling methods.Other approaches,such as SONAR,ultrasonics,and ground penetrating radar do produce image-like data,but of a too specialized nature to consider here.

The analysis of buried sewage pipes possesses certain unique aspects which in?uence the associated computer vision strategy: Lighting:The pipes are buried,dark and,depending on the modality of imaging,there may be constraints on the lighting possible,particularly in the case of CCTV imaging.

Patterned background and contrast:Sewage pipes suffer from signi?cant degrees of deposits and staining,which may be dark,affecting image contrast,or may be highly and irregularly pat-terned,looking very much like any of a number of sewage fail-ure classes –holes,single cracks,networks of cracks,root intrusion,etc.

Limited quality and quantity of data:The slow,expensive approach to data collection strongly limits the total amount of data available for machine learning.Furthermore the lack of standardization –varied methodologies of imaging,machine standards,concrete pipe standards,concrete pipe contents and staining –make it challenging to learn broadly applicable approaches.

The methods of image analysis in the literature mostly involve feature extraction or modeling,both of which are widely used in computer vision and machine learning.Feature extraction [80]is the crucial bridge between a raw image and an information-rich feature vector that can be used for classi?cation.The related prob-lems of image modeling fall into three categories in the context of pipe inspection,from the most speci?c to the most abstract:of parametric/explicit models,morphology/shape-based models,and implicit/black-box models.

3.3.1.Feature extraction

Methods of pattern recognition and classi?cation,such as a sup-port vector machine or nearest neighbor classi?er [80],expect to be given a vector of values describing the object to be classi?ed.An image,containing thousands to millions of pixels,represents data in far too dilute a form to be classi?ed,since computation time and training data requirements are exponential in the number of dimensions.Feature extraction is essentially dimensionality reduction;in the context of analyzing images,computer vision has developed a vast range of approaches for extracting salient features.

Because buried concrete pipes are patterned and poorly lit,robust feature extraction is an essential step and appears through-out the pipe inspection literature.Methods include edge detection [25,22]or the Hough transform [22]for edge/line detection,image

segmentation [26]and background subtraction [18]for foreground object extraction,methods of image registration [18]and optical ?ow [24]for the tracking and association of objects in successive video frames,particularly relevant in CCTV imaging.More advanced methods include texture-based methods,including co-occurrence [21]and histograms of oriented gradients [23],and multi-resolution or wavelet-based approaches [29,17].Not all of these methods can be described here,and the reader is referred to a comprehensive review [81].

3.3.2.Parametric models

In principle,any object which we can recognize in an image,such as a crack,hole,or joint,can be modeled parametrically,with parameters explicitly describing properties such as width,length,radius,and color.The strength of parametric models lies in their explicit nature,being relatively easy to understand and diagnose,however their limitation lies in their limited generalizability:in practice,any special case for which a given model is unprepared leads to a further iteration with a newly revised model addressing that case,and after repeated such iterations leading to ugly,clunky models containing a variety of exceptions.

Given an explicit model,the most fundamental,albeit slow,approach to detecting such objects in an image is using a general-ized Hough transform [82,83].Essentially the Hough transform is a matched ?lter,placing the model in all possible parametric permu-tations at all points in the image and asking regarding degree of ?t.If the number of parameters is suf?ciently few,say two parameters describing the position plus one or two parameters describing size and shape,then the Hough approach may be possible,but given ?ve or more parameters the Hough search space becomes far too large to search densely,and optimization approaches are needed.Signi?cant challenges for parametric approaches arise,by de?nition,for those objects which cannot be well modeled.So whereas a joint (line)or lateral (circle)is relatively simple,a crack is more challenging but may be modeled as a set of connected line segments,but a model to describe the wide range of appearances of root intrusions is very dif?cult.Most parametric computer vision models focus on crack detection,such as modeling a crack as being darker or having a higher variance than its immediate surround-ings [25]or as a set of segments [22].

3.3.3.Morphology

Image morphology represents image shape on the basis of mathematical operations such as shape erosion (shrinking)and dilation (growing).The morphological approaches are more limit-ed than parametric ones since,in principle,a parametric model can encode any imaginable behavior,however the strength of mor-phological approaches is their elegance and operating in a manner similar to humans.

Any morphological operation is described or controlled through a structuring element,normally a relatively simple shape,such as a line,a rectangle,or a disc,which controls the extent to which a given pixel in the image affects its neighbors in dilating or eroding.Many textbooks and tutorial papers have been written [84,85]and the interested reader is referred to them for greater background.Much of pipe inspection is on the basis of binary (light/dark)primitive shapes,making image morphology a natural tool.The most basic shapes are elongated (cracks,joints)and round (holes,laterals),and so analysis can proceed on the basis of one or more round and one or more rectangular structuring elements.Recent uses of morphological approaches in buried pipes can be found in Sinha et al.[16],Su et al.[86],and Halfawy et al.[22].

3.3.

4.Neural models

There has been a huge resurgence in computer vision interest in neural-like models,particularly in the area of deep belief networks

C.Koch et al./Advanced Engineering Informatics xxx (2015)xxx–xxx

[87].The key advantage of a neural approach is that all stages of the problem –contrast enhancement,feature extraction,texture/shape analysis,classi?cation –are machine learned all at once,in an integrated fashion.If the machine learning optimization con-verges well,then the integrated approach can offer robust classi?cation.

On the other hand the sewage pipe problem,with huge num-bers of images and a wide range of background patterning and tex-ture,is a very large nonlinear optimization problem for which convergence may be poor.Neural-like methods are essentially black-box in nature,and therefore the actual effect or role of indi-vidual parameters is exceptionally hard to understand,in contrast to parametric models where the researcher can understand the operations of different parts of the algorithm and where,although parameters would ideally be machine learned,in principle the parameters could be tuned by hand on the basis of an understand-ing of their effect.

Nevertheless,the limitations of the preceding paragraph notwithstanding,neural approaches have seen rather signi?cant application in buried pipe inspection.In most cases,the neural net-work is preceded by computer vision approaches for feature extraction,followed by neural learning [29,27,20,21]or neuro-fuzzy approaches [26,88].

3.3.5.3D reconstruction

A ?nal contribution from computer vision relates to the three dimensional reconstruction of a buried pipe,as a direct geometric detection of deep cracks and holes,rather than indirectly through visual appearance.The computer vision literature has developed a vast range of methods for 3D reconstruction,most notably shape from shading and stereo vision,both relatively complex problems.In contrast,the instruments for pipe inspection employ a laser and generate 3D shape one dot at a time,a far more constrained prob-lem and relatively simple compared to 3D scene reconstruction from images.

The use of laser reconstruction is widespread in computer vision,to generate 3D models of heads,limbs for prosthetics,or objects for 3D printing.For pipe inspection,methods for 3D recon-struction based on laser illumination are developed in Duran et al.[27,20]and Kawasue et al.[89].3.4.Asphalt pavements

3.4.1.Pre-processing

differently well according to varying lighting conditions and shad-ows.Fig.2illustrates the so-called checker shadow illusion [92].Square A looks darker than square B,but their pixel intensities are equal.This means,humans might be able to easily identify an asphalt crack in an image because it appears darker compared to the local https://www.wendangku.net/doc/e711256220.html,puters,however,may fail as they sometimes solely rely on global intensity values.

Several solutions to the non-uniform lighting problem have been proposed.Varadharajan et al.[93]select only images that were taken during daytime and when the weather was overcast or mostly cloudy,so that the lighting conditions are good.The dis-advantage of this approach is that the selection process is also time-consuming and all captured images must be saved before selection and processing,which results in large amounts of data that is stored.Cheng [94]proposed a method to convert all images to a standardized background.For that purpose,a frame is split into rectangular windows.The average light intensity of the pixels in the windows is calculated for each window.Notably low average values are then replaced by the average value of the neighbor win-dows.Finally,multipliers are generated based on the average val-ues.The multipliers are interpolated for each pixel so that all intensities vary around a base intensity.Zou [95]proposed a geo-desic shadow-removal algorithm to remove the pavement shad-ows while preserving the cracks in images.

Another issue related to distress detection in pavement images is the presence of lane-marking on the images.Nguyen et al.[96]detect lane-marking regions and do not consider these regions for the distress detection.First,a binary image is obtained by applying a threshold.Second,the probabilistic Hough Transform is used to detect lines on this binary https://www.wendangku.net/doc/e711256220.html,ne-markings are detected based on the orientations and dimensions of these lines.A range of techniques are applied to eliminate noise or for image enhancement.Lokeshwor [97]and Radopoulou [98]use median ?ltering and morphological operations (erosion,dilation,opening,closing).Li [99]applies Gaussian smoothing for further denoising.Varadharajan [93]calculates the blur magnitude in the images and considers for assessment only images for which the blur-score is below a certain threshold.In some cases it might also be bene?cial to compress the images to reduce the size and computation time,as done by Salman [100].

3.4.2.Defect detection

Several methods have been proposed,which are capable of detecting different types of distress in pavement images.Zhou checker shadow illusion [88]:The squares marked A and B share the same grey intensity (ó1995,Edward 10 C.Koch et al./Advanced Engineering Informatics xxx (2015)xxx–xxx

on the area covered by the distress pixels categorizes video frames as frames with distress or frames without distress.Most detection methods are developed for a speci?c type of distress.Some of the methods are presented below.

3.4.2.1.Cracks.As cracks are the most common distress type,a plenty of crack detection algorithms have been developed and pre-sented.In particular,methods for real time crack analysis [103,104],crack classi?cation [105]crack depth estimation from vision [106],and automating crack sealing have been presented [107,108].

Most of the algorithms for crack detection are based on the assumption that crack pixels are darker than the surroundings.Based on statistical measures of the pixel intensities,thresholding methods that classify pixels as crack or non-crack pixels are applied.Tsai et al.[91]have made a critical assessment of distress segmentation methods,in particular statistical thresholding,Can-ny edge detection,multiscale wavelets,crack seed veri?cation,iterative clipping methods,and dynamic optimization based meth-ods.Koutsopoulos et al.[109]developed an algorithm for crack image segmentation based on a model that describes the statistical properties of pavement images.

Huang et al.[104]also proposed a classi?cation method.An image is divided into cells.Depending on the contrast of each cell to its neighbor,the cells are classi?ed as crack or non-crack cells.However,a limitation of the method is that it is hard to ?nd a uni-versal contrast threshold [91].

Salman et al.[100]proposed an algorithm which uses a Gabor ?lter.The preprocessed pavement image is convolved with the ?l-ter and the real component of the result image is thresholded to generate the binary image.Binary images resulting from different-ly oriented ?lters are combined and an output image is produced.The output image contains detected crack segments.

Moussa and Hussain [110]presented an approach for automatic crack detection,classi?cation and parameter estimation based on machine learning.They apply Graph Cut segmentation to segment an image into crack and background pixels.A binary vector is created after segmentation.Seven features are extracted from the vector for classi?cation purposes.Then,a Support Vector Machine is used to classify the crack type in transverse cracking,longitudinal cracking,block cracking or alligator cracking.Moussa and Hussain also pre-sented an approach to compute the crack extent and severity based on the length and the width of the crack in the image [110].

Varadharajan et al.[93]also use machine learning.They assume input images which can contain background,such as cars,traf?c signs and buildings.First,the ground plane is segmented out from the rest of the image.After that,feature descriptors are computed based on the color and texture of the preprocessed pixels.A total of nine features and data obtained from human annotators are used to train a Support Vector Machine which classi?es the images.Li et al.[99]partition the image into crack regions and regions without cracks using the difference value between the maximum and the minimum grayscales of an image region.Then,the fore-ground is separated from the background by segmenting with Otsu’s method and the images are classi?ed using binary trees and back propagation neural networks.

Zou et al.[95]analyze the intensity difference in regions of the image to determine whether the pixels belong to cracks or not.After that,using tensor voting,a crack map is produced.In the crack map the probability of the pixels that are likely to be located along long crack curves is enhanced.The cracks in the image may sometimes be disconnected,so Zou et al.connect the crack parts with the help of an edge pruning algorithm.

https://www.wendangku.net/doc/e711256220.html,ually,potholes also differ signi?cantly from the background surface.Current computer vision research efforts in

automating the detection of potholes can be divided into 3D recon-struction-based,2D vision-based methods.Detection methods that are based on a 3D reconstruction of the pavement surface rely on 3D point clouds provided by stereo-vision algorithms using a pair of video cameras.Also there are hybrid systems available that use digital cameras to capture consecutive images of lines projected by infrared lasers [111].A stereo-vision based surface model for com-prehensive pavement conditioning has been proposed by Wang [112]and Hou et al.[113].With the availability of a 3D point cloud,Chang et al.[114]have presented a clustering approach that can quantitate the severity and coverage of potholes and Jiaqiu et al.[115]have created a method for identifying,locating,classifying and measuring sag deformations like potholes and depression.The drawbacks of stereo-vision-based approaches are that they require a complete 3D reconstruction of the pavement surface and that the procedure of matching points between the two views is quite challenging due to the very irregular texture and color of the pavement surface.

Karuppuswamy et al.[116]integrated a vision and motion sys-tem to detect simulated potholes.Their approach detects potholes in the center of a lane.However,it relies on computer generated (simulated)potholes that are larger than 2ft in diameter and white in color.The latter are simpli?ed assumptions that do not re?ect realistic pavement conditions.Jahanshahi et al.[117]used a depth sensor to detect and quantify defects in pavements.Based on the depth values of the pixels,pixels are classi?ed as deep or ?at using thresholding.Then,the maximum depth of the defective regions is computed.However,the limitation of the proposed approach is that the data acquisition system,which is the Kinect sensor,is designed for indoor use.As a result,all the captured depth values are zero when the Kinect is exposed to direct sunlight.

Koch et al.[118]also presented a computer vision based approach for pothole detection in asphalt images.Based on sur-rounding shadows,elliptic shape and grain surface texture,the method identi?es potholes in images.Image segmentation,shape approximation,and texture comparison are performed in this order.The image is divided into defect and non-defect pavement regions using histogram shape based thresholding and the triangle algorithm proposed by Zack et al.[119].The shape of the pothole is approximated by applying morphological thinning and elliptic regression.Finally,the surface texture of the pothole candidate region is compared to the non-defect pavement region using spot ?lter responses.The region is determined as a pothole if the region inside the pothole candidate is coarser and grainer than the one outside.Koch et al.extended the method with video processing [120].Using the described pothole detection method,potholes in a sequence of pavement images are counted.

3.4.2.3.Patches.Ca?so et al.[121]observed that pixels which belong to patches have different gray levels from the pixels which belong to the background.They use a clustering method to analyze the image with respect to patches.

Radopoulou et al.[98]detect patches in pavement images by applying morphological operations.Patch regions are segmented based on the assumption that patch pixels have greater intensities than pixels belonging to the background.Then,texture information is utilized and four different ?lters are applied.Subsequently,fea-ture vectors of both intact and patch regions are constructed and compared after the convolution of the image with the ?lters.4.Achievements and challenges

This section summarizes the current achievement and open challenges of computer vision for infrastructure condition assess-ment.A corresponding overview regarding the level of automation

C.Koch et al./Advanced Engineering Informatics xxx (2015)xxx–xxx

in defect detection and condition assessment is presented in Table 5.

4.1.Achievements

When looking at defect detection and condition assessment of reinforced concrete bridges –classi?ed as both vertical and horizontal civil infrastructure –it can be concluded that the cur-rent state-of-the-art computer vision based methods contribute successfully to the automation of detection and measurements of defects.The detection,localization and properties retrieval of both concrete cracks and concrete spalling is to a very large degree auto-mated.Spalling defects can even be quanti?ed and to some extend be mapped to condition ratings.Other important achievements include the ability of computer vision based methods to successful-ly support the detection of connectivity losses between composite sections,changes in boundary conditions,changes in substructure settlements and de?ection of structural members.The accuracy of vision based de?ection detection can even compete with methods employing high accurate laser scanners.

With regard to very long horizontal civil infrastructure,such as precast concrete tunnels,underground concrete pipes and asphalt road networks,it is found that respective data collection technolo-gies are fully automated.Moreover,available computer vision based algorithms successfully support the automation of detecting and localizing defects,such as cracks and joint spalling in concrete tunnels;cracks,holes and joint damage in concrete pipes;and cracks,potholes and patches in asphalt pavements.In case of bridge and tunnel inspection,computer vision based visualization methods (e.g.image stitching)successfully assist in defect detec-tion and assessment as they improve the defect detection results due to better resolution.Concerning asphalt pavements,the crack properties retrieval procedure (type,with,length)is fully automat-ed and some computer vision based distress quanti?cation mea-sures have the potential to be converted to indexes for distress assessment.4.2.Challenges

Concerning computer vision supported concrete bridge inspec-tion,it has to be mentioned that the process of image and video data collection is not yet fully automated.In terms of crack detec-tion and assessment,existing methods need to be improved as performances on noisy data are questionable and accuracies vary with camera pose,camera distance and environmental conditions (lighting and shading at different locations).Moreover,several methods still require a signi?cant amount of manual user input.In general,most of the methods assume images from simple ?at and curved concrete surfaces,so that they may fail in cases of more complex geometries and material,such as joints,seals and bear-ings.Accordingly,there are currently no methods available that support the detection and assessment of bearing distortion and misalignment.

When looking at underground civil infrastructure,such as tun-nels and pipes,it is concluded that poor lighting conditions,irregularly patterned background and contrast as well as limited data quality and quantity impose the most signi?cant problems when dealing with computer vision based approaches to defect detection and assessment.With respect to lighting,common meth-ods either use prior knowledge,thus can hardly be generalized or they rely on some degree of manual input and therefore do not scale well.More recent methods that use machine learning strong-ly rely on training data to create robust classi?https://www.wendangku.net/doc/e711256220.html,ually,the training process is based on supervised learning concepts (manual labeling)and is therefore labor-intensive and error prone.With regard to pipe inspection,the limited amount of data for machine learning and the lack of standardization on defect patterns prevent those methods to perform reasonably well.In addition,detection models with few parameters have limited generalizability,where-as models with many parameters fail in environments with a wide range of background pattern and texture due to the poor conver-gence of inherent non-linear optimization problems.

With respect to asphalt pavement monitoring,natural weather conditions and the daytime determine the success of available computer vision based defect detection and assessment methods.Shadows from trees,for example are very natural and prevent sev-eral methods,which usually work well in good lighting conditions,to perform reasonably well in real environments.Moreover,many algorithms endeavor to perform real-time and therefore are based on some kind of thresholding.However,these methods are not robust enough for image data with average image quality in prac-tice as it is hard to ?nd universal thresholds.Consequently,fully automated and comprehensive pavement distress detection and classi?cation in a real-time environment has remained a challenge.Also,there is no comprehensive and robust method available to determine the severity level of distress for defect and condition assessment of asphalt pavements.

In general,reliable defect detection and condition assessment of civil infrastructure must be based not only on visual inspection methods.First,computer vision methods work under the principle ‘‘What you see is what you can analyze.’’This means,that scenes under observation have to be suf?ciently illuminated to make computer vision methods work.Visible shadows,for example,might have a signi?cant impact on the capability of CV methods.In case of pothole detection shadows support the process,where in cases of 3D reconstruction they hinder the procedure.Moreover,the internal condition of infrastructure components cannot be cap-tured,thus neither assessed using visual methods.On top of visual assessment techniques (whether manual or CV-supported),other advanced in-depth inspection methods (so-called Non-destructive evaluation (NDE)methods)are required to assess the overall con-dition,such as sonic,ultrasonic,magnetic,electrical,nuclear,ther-mography,radar technologies.However,defects on the surface are good indicators of the overall condition as they are part of many visual condition assessment manuals.Second,the data quality plays an important role in terms of noise,distance and perspective to the object of interest and the corresponding image resolution.For instance,if one wants to detect a crack of 1mm width,he or she has to make sure that this 1mm is mapped to a least 1image pixel.Third,a number of safety risks are associated with working at certain heights and under heavy traf?c.In this case,however,emerging remote-controlled unmanned aerial vehicles (UAV)might be a good practical solution for this issue.Forth,the opera-tion of cameras always has to face privacy issues when monitoring public scenes,such as bridges and roads.Thus,it is recommended avoiding people in image and video data.

In summary,the authors conclude that more studies need to be conducted to improve the methods and algorithms for integrated condition assessment.It is currently not possible to detect,mea-sure assess and document all different defects as independent enti-ties to provide an integrated and comprehensive approach for bridge,tunnel,pipe and asphalt inspections.This is mainly due to the unsolved problem of identifying and assessing multiple interacting defects at the same location and the lack of standard-ization in identifying relevant defect parameters to comprehen-sively represent defect information.Moreover,no publically available large datasets exist to leverage supervised learning meth-ods for the robust detection and classi?cation of several infrastruc-ture defect types.

The following listing highlights the key research questions that have to be addressed by future research both in the civil engineer-ing and computer science community in order to take the quality

12 C.Koch et al./Advanced Engineering Informatics xxx (2015)

xxx–xxx

of computer vision based defect detection and condition assess-ment of civil infrastructure to the next level:

How can we comprehensively detect,measure and assess inter-acting defect patterns at the same location to support integrat-ed condition assessment of civil infrastructure?

How can we generalize available detection models to adequate-ly and universally address realistic environmental conditions,such as noisy image and video data,varying lighting conditions,different surface geometries and materials,and different cam-era poses and distances?

How can we limit the amount of manual user input to improve the level of automation from poor defect detection to sophisti-cated defect and condition assessment?

How can we create suf?ciently large,publically available and standardized datasets to leverage the power of existing super-vised machine learning methods for detection,classi?cation and assessment of defects?

How can we create unsupervised machine learning methods (online learning)for ef?cient training and on-demand updating of model parameters in defect detection and assessment models?5.Summary

To ensure the safety and serviceability of civil infrastructure it is essential to visually inspect and assess its physical and functional condition,either at regular intervals (routine inspection)or after disasters (post-disaster inspection).Typically,such condition assessment procedures are performed manually by certi?ed inspectors and/or structural engineers.This process includes the detection of the defects and damage (cracking,spalling,defective joints,corrosion,potholes,etc.)existing on civil infrastructure ele-ments,such as buildings,bridges,roads,pipes and tunnels,and the defects’magnitude (number,width,length,etc.).The condition assessment results are used to predict future conditions,to support investment planning,and to allocate limited maintenance and repair resources.

This paper has presented the current practices of assessing the visual condition of vertical and horizontal civil infrastructure,in particular of reinforced concrete bridges (horizontal:decks,gird-ers,vertical:columns),precast concrete tunnels (horizontal:seg-mental lining),underground concrete pipes (horizontal)(wastewater infrastructure),and asphalt pavements (horizontal).Following this,the second and largest part of the paper has focused

on a comprehensive synthesis of the state of the art in computer vision based defect detection and condition assessment of civil infrastructure.Several methodologies have been described and categorized,and literature on respective tests and evaluations on the current performances to detect and measure different defect and damage pattern in remote and close-up images of buildings,bridges,roads,pipes and tunnels has been presented.In the third part of this paper the current achievements and limitations of com-puter vision for infrastructure condition assessment have been summarized.Finally,open research challenges have been outlined to assist both the civil engineering and the computer science research community in setting an agenda for future research.

References

[1]National Transportation Safety Board,Collapse of I-35W Highway Bridge,

Minneapolis,Minnesota,August 1,2007,2008.

[2]T.Asakura,Y.Kojima,Tunnel maintenance in Japan 18(2–3)(2003)161–169.[3]R.Zaurin,F.N.Catbas,Integration of computer imaging and sensor data for

912structural health monitoring of bridges,Smart Mater.Struct.19(1)(2010).

[4]AASHTO Publication,Manual for Bridge Element Inspection,2013.

[5]Z.Zhu,S.German,I.Brilakis,Detection of large-scale concrete columns for

automated bridge inspection,Autom.Construct.19(8)(2010)1047–1055.[6]FHWA Report,Reliability of Visual Inspection for Highway Bridges,2001.[7]NHI-FHWA Online Course,Introduction to Safety Inspection of In-Service

Bridges.

[8]Federal Highway Admistration (FHWA),Tunnel Operations,Maintenance,

Inspection and Evaluation (TOMIE)Manual,2011.

[9]Federal Highway Administration (FHWA)and Federal Transit Admistration,

Highway and Rail Transit Tunnel Inspection Manual,U.S.Department of Transportation,2005.

[10]National Cooperative Highway Research Program Project 20-68A,Best

Practices for Roadway Tunnel Design,Construction,Maintenance,Inspection,and Operation,2011.

[11]Pavemetrics,Laser tunnel scanning system (LTSS),Advanced Vision

Technology 3D |Pavemetrics TM,Pavemetrics Systems Inc.,(accessed 03.11.14).

[12]DIBIT,Tunnel Scanning,Tunnel,Service,Home,DIBIT Messtechnik GmbH,

DIBIT Messtechnik GmbH (accessed 03.11.14).

[13]B.Cohen,Fixing America’s crumbling underground water infrastructure,

Issue Anal.4(2012).

[14]T.Iseley,D.M.Abraham S.Gokhale,Intelligent sewer condition evaluation

technologies,in:Proceedings of the International NO-DIG Conference,1997,pp.254–265.

[15]R.Wirahadikusumah, D.M.Abraham,T.Iseley,R.K.Prasanth,Assessment

technologies for sewer system rehabilitation,J.Autom.Construct.7(4)(1998)259–270.

[16]S.K.Sinha,P.Fieguth,Segmentation of buried concrete pipe images,Autom.

Construct.15(2006)47–57.

[17]O.Duran,K.Althoefer,L.Seneviratne,State of the art in sensor technologies

for sewer inspection,IEEE Senors J.2(2)(2002)73–81.

Table 5

Level of automation in computer vision based defect detection and condition assessment:(+)achieved,($)partially achieved,(à)not achieved

yet.

C.Koch et al./Advanced Engineering Informatics xxx (2015)xxx–xxx

化学专业英语(修订版)翻译

01 THE ELEMENTS AND THE PERIODIC TABLE 01 元素和元素周期表 The number of protons in the nucleus of an atom is referred to as the atomic number, or proton number, Z. The number of electrons in an electrically neutral atom is also equal to the atomic number, Z. The total mass of an atom is determined very nearly by the total number of protons and neutrons in its nucleus. This total is called the mass number, A. The number of neutrons in an atom, the neutron number, is given by the quantity A-Z. 质子的数量在一个原子的核被称为原子序数,或质子数、周淑金、电子的数量在一个电中性原子也等于原子序数松山机场的总质量的原子做出很近的总数的质子和中子在它的核心。这个总数被称为大量胡逸舟、中子的数量在一个原子,中子数,给出了a - z的数量。 The term element refers to, a pure substance with atoms all of a single kind. T o the chemist the "kind" of atom is specified by its atomic number, since this is the property that determines its chemical behavior. At present all the atoms from Z = 1 to Z = 107 are known; there are 107 chemical elements. Each chemical element has been given a name and a distinctive symbol. For most elements the symbol is simply the abbreviated form of the English name consisting of one or two letters, for example: 这个术语是指元素,一个纯物质与原子组成一个单一的善良。在药房“客气”原子的原子数来确定它,因为它的性质是决定其化学行为。目前所有原子和Z = 1 a到Z = 107是知道的;有107种化学元素。每一种化学元素起了一个名字和独特的象征。对于大多数元素都仅仅是一个象征的英文名称缩写形式,一个或两个字母组成,例如: oxygen==O nitrogen == N neon==Ne magnesium == Mg

应用化学专业英语翻译完整篇

1 Unit5元素周期表 As our picture of the atom becomes more detailed 随着我们对原子的描述越来越详尽，我们发现我们陷入了进退两难之境。有超过100多中元素要处理，我们怎么能记的住所有的信息？有一种方法就是使用元素周期表。这个周期表包含元素的所有信息。它记录了元素中所含的质子数和电子数，它能让我们算出大多数元素的同位素的中子数。它甚至有各个元素原子的电子怎么排列。最神奇的是，周期表是在人们不知道原子中存在质子、中子和电子的情况下发明的。Not long after Dalton presented his model for atom( )在道尔顿提出他的原子模型（原子是是一个不可分割的粒子，其质量决定了它的身份）不久，化学家门开始根据原子的质量将原子列表。在制定像这些元素表时候，他们观察到在元素中的格局分布。例如，人们可以清楚的看到在具体间隔的元素有着相似的性质。在当时知道的大约60种元素中，第二个和第九个表现出相似的性质，第三个和第十个，第四个和第十一个等都具有相似的性质。 In 1869,Dmitri Ivanovich Mendeleev,a Russian chemist, 在1869年，Dmitri Ivanovich Mendeleev ,一个俄罗斯的化学家，发表了他的元素周期表。Mendeleev通过考虑原子重量和元素的某些特性的周期性准备了他的周期表。这些元素的排列顺序先是按原子质量的增加,,一些情况中, Mendeleev把稍微重写的元素放在轻的那个前面.他这样做只是为了同一列中的元素能具有相似的性质.例如,他把碲(原子质量为128)防在碘(原子质量为127)前面因为碲性质上和硫磺和硒相似, 而碘和氯和溴相似. Mendeleev left a number of gaps in his table.Instead of Mendeleev在他的周期表中留下了一些空白。他非但没有将那些空白看成是缺憾，反而大胆的预测还存在着仍未被发现的元素。更进一步，他甚至预测出那些一些缺失元素的性质出来。在接下来的几年里，随着新元素的发现，里面的许多空格都被填满。这些性质也和Mendeleev所预测的极为接近。这巨大创新的预计值导致了Mendeleev的周期表为人们所接受。 It is known that properties of an element depend mainly on the number of electrons in the outermost energy level of the atoms of the element. 我们现在所知道的元素的性质主要取决于元素原子最外层能量能级的电子数。钠原子最外层能量能级（第三层）有一个电子，锂原子最外层能量能级（第二层）有一个电子。钠和锂的化学性质相似。氦原子和氖原子外层能级上是满的，这两种都是惰性气体，也就是他们不容易进行化学反应。很明显，有着相同电子结构（电子分布）的元素的不仅有着相似的化学性质，而且某些结构也表现比其他元素稳定（不那么活泼） In Mendeleev’s table,the elements were arranged by atomic weights for 在Mendeleev的表中，元素大部分是按照原子数来排列的，这个排列揭示了化学性质的周期性。因为电子数决定元素的化学性质，电子数也应该（现在也确实）决定周期表的顺序。在现代的周期表中，元素是根据原子质量来排列的。记住，这个数字表示了在元素的中性原子中的质子数和电子数。现在的周期表是按照原子数的递增排列，Mendeleev的周期表是按照原子质量的递增排列，彼此平行是由于原子量的增加。只有在一些情况下（Mendeleev注释的那样）重量和顺序不符合。因为原子质量是质子和中子质量的加和，故原子量并不完全随原子序数的增加而增加。原子序数低的原子的中子数有可能比原子序数高的原

当今最潮的英语翻译

伪球迷biased fans 紧身服straitjacket 团购group buying 奉子成婚shortgun marriage 婚前性行为premartial sex 开博to open a blog 家庭暴力family volience 问题家具problem furniture 炫富flaunt wealth 决堤breaching of the dike 上市list share 赌球soccer gambling 桑拿天sauna weather 自杀Dutch act 假发票fake invoice 落后产能outdated capacity 二房东middleman landlord 入园难kindergarten crunch 生态补偿ecological compensation 金砖四国BRIC countries 笑料laughing stock 泰国香米Thai fragrant rice 学历造假fabricate academic credentials 泄洪release flood waters 狂热的gaga eg: I was gaga over his deep blue eyes when I first set eyes on him 防暑降温补贴high temperature subsidy 暗淡前景bleak prospects 文艺爱情片chick flick 惊悚电影slasher flick 房奴车奴mortgage slave 上课开小差zone out 万事通know-it-all 毕业典礼commencement 散伙饭farewell dinner 毕业旅行after-graduation trip 节能高效的fuel-efficient 具有时效性的time-efficient 死记硬背cramming 很想赢be hungry for success 面子工程face job 捉迷藏play tag 射手榜top-scorer list 学历门槛academic threshold 女学究blue stocking

漂亮的英文翻译

漂亮的英文翻译 1.你可知我百年的孤寂只为你一人守侯，千夜的恋歌只为你一人而唱。 You know my loneliness is only kept for you, my sweet songs are only sung for you. 2.如果活着，是上帝赋予我最大的使命，那么活者有你，将会是上帝赋予我使命的恩赐…… If living on the earth is a mission from the lord… living with you is the award of the lord… 3.你知道思念一个人的滋味吗，就像喝了一大杯冰水，然后用很长很长的时间流成热泪。 Do you understand the feeling of missing someone? It is just like that you will spend a long hard time to turn the ice-cold water you have drunk into tears.

4.在这充满温馨的季节里，给你我真挚的祝福及深深的思念。 In such a soft and warm season, please accept my sincere blessing and deep concern for you. 5.一份不渝的友谊，执着千万个祝福，给我想念的朋友，温馨的问候。 For our ever-lasting friendship, send sincere blessings and warm greetings to my friends whom I miss so much. 6.想你，是一种美丽的忧伤的甜蜜的惆怅，心里面，却是一种用任何语言也无法表达的温馨。 It is graceful grief and sweet sadness to think of you, but in my heart, there is a kind of soft warmth that can’t be expressed with any choice of words. 7.不同的时间，不同的地点，不同的人群，相同的只有你和我；时间在变，空间在变，不变的只有对你无限的思念！ You and I remains the same in different time, at different places,among different people; time is changing, space is changing and everything is changing except my miss to you! 8.没有杯子……咖啡是寂寞的……没有你……我是孤独的…… Coffee is lonely without cups. I am lonely without you. 9.每一天都为你心跳，每一刻都被你感动，每一秒都为你担心。有你的感觉真好。

化学专业英语翻译1

01.THE ELEMENTS AND THE PERIODIC TABLE 01元素和元素周期表。 The number of protons in the nucleus of an atom is referred to as the atomic number, or proton number, Z. The number of electrons in an electrically neutral atom is also equal to the atomic number, Z. The total mass of an atom is determined very nearly by the total number of protons and neutrons in its nucleus. This total is called the mass number, A. The number of neutrons in an atom, the neutron number, is given by the quantity A-Z. 原子核中的质子数的原子称为原子序数，或质子数，卓电子数的电中性的原子也等于原子序数Z，总质量的原子是非常接近的总数量的质子和中子在原子核。这被称为质量数，这个数的原子中的中子，中子数，给出了所有的数量 The term element refers to, a pure substance with atoms all of a single kind. To the chemist the "kind" of atom is specified by its atomic number, since this is the property that determines its chemical behavior. At present all the atoms from Z = 1 to Z = 107 are known; there are 107 chemical elements. Each chemical element has been given a name and a distinctive symbol. For most elements the symbol is simply the abbreviated form of

应用化学专业英语第二版万有志主编版课后答案和课文翻译

Uｎit 1 The RｏotｓoｆCｈemisｔry Ｉ．Ｃoｍpreｈensioｎ. 1。Ｃ 2. Ｂ3．D 4. C 5. B ＩI。Ｍａｋe ａsenｔeｎｃe ｏut of ｅach ｉｔem ｂy reaｒranging ｔhe wordｓin ｂracｋets. 1.Tｈｅｐuｒiｆｉcation of aｎorganｉc compounｄis usually a matter of consｉｄerａｂlｅdｉfｆicｕｌtｙ, and iｔis nｅcｅｓsaｒy to eｍploy vａrious methoｄs fｏr thｉｓpurpｏse。 2.Science is an ｅｖeｒ－iｎcｒｅasing boｄy ofａccｕｍｕlａｔed anｄｓｙsｔｅmatiｚed knowledｇe ａnd iｓalsｏan activｉty bｙwｈic ｈｋnowledge iｓｇenerated。 3.Lifｅ，aftｅr all, is onｌy cｈeｍistry,in faｃｔ, a sｍall example of ｃ hｅｍistry ｏbsｅｒved oｎａsｉngle munｄane pｌａnet。 4.Peoplｅaｒe made of molｅculｅｓ; sｏmｅｏf ｔhｅmｏleculｅs in p ｅople aｒe rａｔher ｓimple whｅreas otｈerｓａｒｅhｉghly compｌex。 5.Chｅｍistry ｉｓever preseｎｔin ｏｕｒｌｉves frｏm ｂｉｒtｈ tｏｄeａthｂeｃａｕse withoｕt chｅmｉｓtrｙｔhere iｓneither life nor deａth. 6.Maｔhｅｍaｔics apｐeａｒs ｔo be almost as humaｎkiｎｄand ａｌｓo ｐermeaｔeｓall aspeｃts of ｈuman life, aｌthough ｍａｎｙof us are ｎｏｔfuｌly ａwａrｅoｆtｈis. III。Tｒanｓlatioｎ. 1.（a)chｅmicaｌprｏcess (b) natuｒal scienｃe（c）the teｃhｎｉｑｕe of distｉllaｔioｎ 2.Iｔｉs theａtｏｍs that ｍａkｅuｐｉron, waｔeｒ，oｘygen and the lｉke/anｄso on／aｎｄsｏfortｈ/aｎd otherｗiｓe. 3.Cheｍisｔｒy haｓa very long hｉｓtorｙ, ｉnｆaｃｔ，human a ｃtivity in chｅｍｉstrｙgｏes bａｃk ｔo ｐreｒecｏrｄed tｉmｅｓ/pｒedatiｎg reｃｏrded times． 4.Aｃcoｒdｉng to/Froｍthe eｖapoｒaｔion oｆwaｔeｒ，pｅople know ／reaｌized ｔｈat ｌiquidｓcａn ｔurn/bｅ/changeｉｎto gases ｕndeｒcertain conditions/circｕmsｔance／enｖironment。 5.Yoｕｍusｔknow ｔhe propｅｒtieｓｏfｔｈe materiaｌbeｆore y ｏu use it. IＶ．Ｔrａnslation 化学是三种基础自然科学之一,另外两种是物理和生物.自从宇宙大爆炸以来，化学过程持续进行,甚至地球上生命的出现可能也是化学过程的结果。人们也许认为生命是三步进化的最终结果,第一步非常快，其余两步相当慢.这三步

50个很潮的英文单词

发表日期：2015-09-29 07:48 来源：80后励志网编辑：80后点击：3321次文章标签: 英语名言教育好文读书励志英语教育文章导读:英语是国际性的语言，英语在我们的生活中使用率也越来越高，下面这50个很潮的英文单词，年轻人一定要学会哦！ 50个很潮的英文单词，年轻人一定要学会！ 1.预约券 reservation ticket 2.下午茶 high tea 3.微博 Microblog/ Tweets 4.裸婚 naked wedding 5.亚健康 sub-health 6.平角裤 boxers 7.愤青 young cynic 8.灵魂伴侣 soul mate 9.小白脸 toy boy 10.精神出轨 soul infidelity 11.人肉搜索 flesh search 12.浪女 dillydally girl 13.公司政治 company politics 14.剩女 3S lady(single，seventies,stuck)/left girls 15.山寨 copycat 16.异地恋 long-distance relationship 17.性感妈妈 yummy mummy ; milf(回复中指出的~) 18.钻石王老五 diamond bachelor;most eligible bachelor 20.时尚达人 fashion icon 21.御宅 otaku 22.上相的,上镜头的 photogenic 23.脑残体 leetspeak 24.学术界 academic circle 25.哈证族 certificate maniac 26.偶像派 idol type 27.住房公积金 housing funds 28.个税起征点 inpidual income tax threshold 29.熟女 cougar(源自电影Cougar Club) 30.挑食者 picky-eater 31.伪球迷 fake fans 32.紧身服 straitjacket 33.团购 group buying 34.奉子成婚 shotgun marriage 35.婚前性行为 premarital sex 36.开博 to open a blog 37.家庭暴力 family/domestic violence (由回复更正) 38.问题家具 problem furniture

漂亮的英文怎么写

漂亮的英文怎么写【漂亮：pretty; beautiful; good-looking; handsome】漂亮[piào liang] (好看;美观) handsome; good-looking; pretty; beautiful: be prettily dressed; 衣服穿得漂亮 trim oneself up; 把自己打扮得漂漂亮亮 This photograph flatters you. 这照片比你本人漂亮。 (出色) smart; remarkable; brilliant; splendid; beautiful: well done; 干得漂亮 speak prettily; 说得漂亮 You write a beautiful hand. 这字写得真漂亮。 pretty例句: 1.Today was a pretty project based intensive day. 今天是项目相当密集的一天。

2.She is a pretty girl. 她是个漂亮的女孩。 3.Morocco's king seems pretty safe. 摩洛哥国王看上去似乎很安全。 4.Life can be pretty complicated. 生活是非常复杂的。 5.Balancing national security concerns against moral responsibilities is never pretty. 平衡国家安全问题和道德责任从来不是一件容易的事。 beautiful例句: 1.But paris is more beautiful than tokyo. 但巴黎是比东京更美丽。 2.You have a beautiful house. 你有个漂亮的房子。 3.There is nothing more beautiful than a wedding. 没有什么比婚礼更美的了。 https://www.wendangku.net/doc/e711256220.html, 4.South africa is a beautiful country. 南非是一个美丽的国度。 5.Then she began a beautiful love song. 接着她开始演唱一首优美的爱情歌曲。

《化学工程与工艺专业英语》课文翻译完整版

Unit 1 Chemical Industry 化学工业 1.Origins of the Chemical Industry Although the use of chemicals dates back to the ancient civilizations, the evolution of what we know as the modern chemical industry started much more recently. It may be considered to have begun during the Industrial Revolution, about 1800, and developed to provide chemicals roe use by other industries. Examples are alkali for soapmaking, bleaching powder for cotton, and silica and sodium carbonate for glassmaking. It will be noted that these are all inorganic chemicals. The organic chemicals industry started in the 1860s with the exploitation of William Henry Perkin‘s discovery if the first synthetic dyestuff—mauve. At the start of the twentieth century the emphasis on research on the applied aspects of chemistry in Germany had paid off handsomely, and by 1914 had resulted in the German chemical industry having 75% of the world market in chemicals. This was based on the discovery of new dyestuffs plus the development of both the contact process for sulphuric acid and the Haber process for ammonia. The later required a major technological breakthrough that of being able to carry out chemical reactions under conditions of very high pressure for the first time. The experience gained with this was to stand Germany in good stead, particularly with the rapidly increased demand for nitrogen-based compounds (ammonium salts for fertilizers and nitric acid for explosives manufacture) with the outbreak of world warⅠin 1914. This initiated profound changes which continued during the inter-war years (1918-1939). 1．化学工业的起源尽管化学品的使用可以追溯到古代文明时代，我们所谓的现代化学工业的发展却是非常近代（才开始的）。可以认为它起源于工业革命其间，大约在1800年，并发展成为为其它工业部门提供化学原料的产业。比如制肥皂所用的碱，棉布生产所用的漂白粉，玻璃制造业所用的硅及Na2CO3. 我们会注意到所有这些都是无机物。有机化学工业的开始是在十九世纪六十年代以William Henry Perkin 发现第一种合成染料—苯胺紫并加以开发利用为标志的。20世纪初，德国花费大量资金用于实用化学方面的重点研究，到1914年，德国的化学工业在世界化学产品市场上占有75%的份额。这要归因于新染料的发现以及硫酸的接触法生产和氨的哈伯生产工艺的发展。而后者需要较大的技术突破使得化学反应第一次可以在非常高的压力条件下进行。这方面所取得的成绩对德国很有帮助。特别是由于1914年第一次世界大仗的爆发，对以氮为基础的化合物的需求飞速增长。这种深刻的改变一直持续到战后（1918-1939）。 date bake to/from: 回溯到 dated: 过时的，陈旧的 stand sb. in good stead: 对。。。很有帮助

应用化学专业英语及答案

黄冈师范学院 2009—2010学年度第一学期期末试卷考试课程：专业英语考核类型：考试A卷考试形式：闭卷出卷教师：杨一思考试专业：化学考试班级：应用化学200601 一、Translate the following into English（20 points） 1.过滤 2.浓缩 3.结晶化 4.吸附 5. 蒸馏6．超临界的 7.二氯甲烷 8.热力学平衡 9.亲电性 10.表面张力 11.共轭的 12.酮 13.平衡常数 14.丙基 15.丁基 16.亚甲基 18.环己酮 19.同位素 20.标准熵二、Translate the following into Chinese（20 points） 1. methyl propanoate 2. rate constant 3. ethyl methyl ketone 4. free energy 5. radical intermediate 6. isobutyl methyl ether 7. 3-chloropropene 8. primary radical 9. n-propyl bromide 10. bond energy 11. circulating electrons 12. local magnetic fields 13. tetramethylsilane 14. mass to charge ratios 15 phenylamine 16 amide 17. amine 18. nucleophile 19. perchlorate 20. carbocation 三、Translation the following into chinese （40 points） A卷【第1页共 3 页】

77个优美英文句子+漂亮翻译

1 我的世界不允许你的消失,不管结局是否完美. No matter the ending is perfect or not, you cannot disappear from my world. 2 爱情是一个精心设计的谎言 Love is a carefully designed lie. 3 承诺常常很像蝴蝶,美丽的飞盘旋然后不见 Promises are often like the butterfly, which disappear after beautiful hover. 4 凋谢是真实的盛开只是一种过去 Fading is true while flowering is past 5 为什么幸福总是擦肩而过,偶尔想你的时候….就让….回忆来陪我. Why I have never catched the happiness? Whenever I want you ,I will be accompanyed by the memory of... 6 如果你为着错过夕阳而哭泣，那么你就要错群星了 If you weeped for the missing sunset,you would miss all the shining stars 7 如果只是遇见，不能停留，不如不遇见 If we can only encounter each other rather than stay with each other,then I wish we had never encountered . 8 宁愿笑著流泪,也不哭著说后悔心碎了,还需再补吗？ I would like weeping with the smile rather than repenting with the cry,when my heart is broken ,is it needed to fix? 9 天空没有翅膀的痕迹，而鸟儿已飞过 There are no trails of the wings in the sky, while the birds has flied away. 10 当香烟爱上火柴时,就注定受到伤害 When a cigarette falls in love with a match,it is destined to be hurt 11 人活着总是要得罪一些人的就要看那些人是否值得得罪 When alive ,we may probably offend some people.However, we must think about whether they are deserved offended. 12 命里有时终需有命里无时莫强求 You will have it if it belongs to you,whereas you don't kveth for it if it doesn't appear in your life.

化工专业英语lesson4翻译汇编

仅供参考 Introduction to Organic Chemistry 1. Sources of Organic Compounds The major sources of organic chemicals are coal, petroleum, and agricultural products. Both coal and petroleum were formed through the geologic processes of changing animal and plant remains into carbon-containing residues. About one-third of all organic chemicals are derived from coal and about one-half from the petroleum industry 有机化合物的来源有机化学药品的主要来源是煤、石油和农产品。动植物的遗体通过地质作用变成含碳残基然后形成煤和石油。三分之一的所有有机化合物品是从煤中得到的，一般来自于石油工业。 2. The Methods and Objectives of Organic Chemistry Because of the tremendous number of organic compounds known, and of the many more being synthesized daily, the study of organic chemistry is not the study of individual compounds, it is the study of groups or families of compounds all closely related to each other. Obviously, the former approach would be prohibitive[prE5hibitiv]. Once the structural relationships of certain typical members of a particular group or family of compounds are understood, these structural features are understood for any one of the many members of the family, even though some may not be known compounds. 因为已知的有机化合物的数目庞大，而且还在逐日合成更多的品种，所以有机化学不是研究单个的化合物，而是把彼此密切相关的化合物按类或族来研究。显然，以前的方法是不可取的，一旦典型的一类特殊化合物被认识，这些结构特征将适用于这类化合物，甚至是一些未知的化合物， For each group or family of compounds often called homologous series of compounds, structural features are important. In studying organic chemistry, it is not enough to know the identities of the elements and how many atoms of each element are present in a given molecule. More importantly, the order in which these atoms are linked together to form

应用化学专业英语翻译(第二版)

Unit10 Nomenclature of Hydrocarbons碳氢化合物的命名 Alkanes烷烃理想的，每一种化合物都应该由一个明确描述它的结构的名称，并且通过这一名称能够画出它的结构式。为了这一目的，全世界的化学家接受了世界纯粹与应用化学会（IUPAC）建立的一系列规则。这个系统就是IUPAC系统，或称为日内瓦系统，因为IUPAC的第一次会议是在瑞士日内瓦召开的。不含支链的烷烃的IUPAC命名包括两部分（1）表明链中碳原子数目的前缀；（2）后缀－ane，表明化合物是烷烃。用于表示1至20个碳原子的前缀见表表中前4个前缀是由IUPAC选择的，因为它们早已在有机化学中确定了。实际上，它们甚至早在它们成为规则之下的结构理论的暗示之前，它们的地位就确定了。例如，在丁酸中出现的前缀but－，一种表示在白脱脂中存在的四个碳原子的化合物（拉丁语butyrum白脱（黄油））。表示5个或更多碳原子的词根来源于希腊或拉丁词根。含取代基的烷烃的IUPAC名称由母体名称和取代基名称组成，母体名称代表化合物的最长碳链，取代基名称代表连接在主链上的基团。来源于烷烃的取代基称为烷基。字母R－被广泛用来表示烷烃的存在.烷烃的命名是去掉原烷基名称中的－ane加上后缀－yl。例如，烷基CH3CH2－称为乙基。 CH3－CH3乙烷（原碳氢化合物）CH3CH2－乙基（一个烷基）下面是IUPAC的烷烃命名规则： 1. 饱和碳氢化合物称为烷烃。 2. 对有支链的碳氢化合物，最长的碳链作为主链，IUPAC命名按此主链命名。 3. 连接在主链上的基团称为取代基。每一取代基有一名称和一数字.这一数字表示取代基连接在主链上的碳原子的位置。 4. 如果有多于一个的相同取代基，要给出表示支链位置的每个数字。而且，表示支链数目的数字由前缀di－，tri－，tetra－，penta－等表示。 5. 如果有一个取代基，主链碳原子编号从靠近支链的一端开始，使支链位号最小。如果有两个或多个取代基，支链从能使第一个取代基位次较小的一侧编号。 6. 如果有两个或多个取代基，它们按字母顺序排列。当排列取代基时，前缀iso－（异）和neo－（新）也按字母排列，前缀sec－（仲）和tert－（特）在字母排列中忽略此外，按字母排列取代基时，表示倍数的前缀2，3，4等也被忽略。 Alkenes烯烃烷烃的碳原子间只有单键。烯烃在两个碳原子间有双键（两个键）。考虑到两个碳原子间以双键相连。因为双键用去了两个碳原子的共4个电子。所以只剩下4个电子以供成

50个很潮的英文单词

50个很潮的英文单词，年轻人一定要学会！发表日期：2015-09-29 07:48 来源：80后励志网编辑：80后点击：3321次文章标签: 英语名言教育好文读书励志英语教育文章导读:英语是国际性的语言，英语在我们的生活中使用率也越来越高，下面这50个很潮的英文单词，年轻人一定要学会哦！ 50个很潮的英文单词，年轻人一定要学会！ 1.预约券 reservation ticket 2.下午茶 high tea 3.微博 Microblog/ Tweets 4.裸婚 naked wedding 5.亚健康 sub-health 6.平角裤 boxers 7.愤青 young cynic 8.灵魂伴侣 soul mate 9.小白脸 toy boy 10.精神出轨 soul infidelity 11.人肉搜索 flesh search 12.浪女 dillydally girl 13.公司政治 company politics 14.剩女 3S lady(single，seventies,stuck)/left girls 15.山寨 copycat 16.异地恋 long-distance relationship 17.性感妈妈 yummy mummy ; milf(回复中指出的~) 18.钻石王老五 diamond bachelor;most eligible bachelor

20.时尚达人 fashion icon 21.御宅 otaku 22.上相的,上镜头的 photogenic 23.脑残体 leetspeak 24.学术界 academic circle 25.哈证族 certificate maniac 26.偶像派 idol type 27.住房公积金 housing funds 28.个税起征点 inpidual income tax threshold 29.熟女 cougar(源自电影Cougar Club) 30.挑食者 picky-eater 31.伪球迷 fake fans 32.紧身服 straitjacket 33.团购 group buying 34.奉子成婚 shotgun marriage 35.婚前性行为 premarital sex 36.开博 to open a blog 37.家庭暴力 family/domestic violence (由回复更正) 38.问题家具 problem furniture 39.炫富 flaunt wealth 40.决堤 breaching of the dike 41.上市 list share 42.赌球 soccer gambling 43.桑拿天 sauna weather 44.自杀 Dutch act 45.假发票 fake invoice 46.落后产能 outdated capacity 47.二房东 middleman landlord 48.入园难 kindergarten crunch 49.生态补偿 ecological compensation 50.金砖四国 BRIC countrie

好漂亮的英文翻译

1.你可知我百年的孤寂只为你一人守侯，千夜的恋歌只为你一人而唱。 You know my loneliness is only kept for you, my sweet songs are only sung for you. 2.如果活着，是上帝赋予我最大的使命，那么活者有你，将会是上帝赋予我使命的恩赐…… If living on the earth is a mission from the lord… living with you is the award of the lord… 3.你知道思念一个人的滋味吗，就像喝了一大杯冰水，然后用很长很长的时间流成热泪。 Do you understand the feeling of missing someone? It is just like that you will spend a long hard time to turn the ice-cold water you have drunk into tears. 4.在这充满温馨的季节里，给你我真挚的祝福及深深的思念。 In such a soft and warm season, please accept my sincere blessing and deep concern for you. 5.一份不渝的友谊，执着千万个祝福，给我想念的朋友，温馨的问候。 For our ever-lasting friendship, send sincere blessings and warm greetings to my friends whom I miss so much. 6.想你，是一种美丽的忧伤的甜蜜的惆怅，心里面，却是一种用任何语言也无法表达的温馨。 It is graceful grief and sweet sadness to think of you, but in my heart, there is a kind of soft warmth that can’t be expressed with any choice of words.

英文翻译-main

化学专业英语(修订版)翻译

应用化学专业英语翻译完整篇

当今最潮的英语翻译

漂亮的英文翻译

化学专业英语翻译1

应用化学专业英语第二版万有志主编版课后答案和课文翻译

50个很潮的英文单词

漂亮的英文怎么写

《化学工程与工艺专业英语》课文翻译 完整版

应用化学专业英语及答案

77个优美英文句子+漂亮翻译

化工专业英语lesson4翻译汇编

应用化学专业英语翻译(第二版)

50个很潮的英文单词

好漂亮的英文翻译

《化学工程与工艺专业英语》课文翻译完整版