CN102183525A - Bridge apparent state automatic detection device and method based on charge coupled device (CCD) array photographic technique - Google Patents
Bridge apparent state automatic detection device and method based on charge coupled device (CCD) array photographic technique Download PDFInfo
- Publication number
- CN102183525A CN102183525A CN2011100230283A CN201110023028A CN102183525A CN 102183525 A CN102183525 A CN 102183525A CN 2011100230283 A CN2011100230283 A CN 2011100230283A CN 201110023028 A CN201110023028 A CN 201110023028A CN 102183525 A CN102183525 A CN 102183525A
- Authority
- CN
- China
- Prior art keywords
- image
- bridge
- ccd
- video camera
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention relates to a bridge apparent state automatic detection device based on a charge coupled device (CCD) array photographic technique and a bridge apparent state automatic detection method based on a charge coupled device (CCD) array photographic technique. The bridge apparent state automatic detection device consists of a CCD video camera array, a scanning frame, an image acquisition card, a computer, detection management software and a detection vehicle and is characterized in that: the CCD video camera array is arranged on the scanning frame; the bottom of the scanning frame is fixed on the detection vehicle positioned on or below a bridge; the CCD video camera array is connected with the image acquisition card through leads; the image acquisition card is arranged on the computer; the CCD video camera array transmits synchronous bridge bottom and side defect images to the computer in real time; and the detection management software realizes synthesis, processing, defect identification and data management of the bridge apparent defect images. A computer video technique is applied to the bridge apparent defect detection and a judgment result is directly provided by a reliable CCD automatic data acquiring and process system, so that time is saved, working efficiency and precision are improved, and the normal operation state of the bridge can be guaranteed.
Description
Technical field
The present invention relates to Automatic Measurement Technique, is exactly apparent condition automatic detection device of a kind of bridge based on the ccd array camera technique and method thereof specifically.
Background technology
Communications and transportation is the economic lifeline of a country, and road and bridge are to make the communications and transportation can straightway carrier, and bridge structure plays crucial effects especially as the transport hub.The a lot of highway bridges of China are in the mountain area or cross over rivers, and along with the increase of active time, bridge structure various fatigues and damage can occur unavoidably, and the defective of most of bridge structures mainly is created in the bridge bottom.For large-scale concrete-bridge, the crack of bridge is to need one of subject matter that detects assessment, and most structural crack all is created in the bridge bottom, makes and detects difficulty especially.According to incompletely statistics, the annual bridge that damages has more than 90% and is caused by the crack.Along with the maximization day by day of bridge structure form, complicated, quality requirements is strict day by day, and the Crack Detection problem of bridge structure becomes the technical barrier with suitable ubiquity.Bridge structure crackle manual detection method inefficiency commonly used at present, and have certain danger.Therefore adopt objective, effective bridge defect detecting technique, the degree of damage evaluation of bridge structure defective is had important significance for theories and practical value.
Summary of the invention
The object of the present invention is to provide a kind of apparent condition automatic detection device of the bridge based on the ccd array camera technique and method thereof of differentiating the result, saving time, improve work efficiency that directly provide by CCD automatic data acquisition disposal system.
The object of the present invention is achieved like this: it is made up of ccd video camera array, scanning frame, image pick-up card, computing machine and inspection vehicle, it is characterized in that: the ccd video camera array is installed on the scanning frame, scanning frame bottom is fixed on the inspection vehicle that is positioned under bridge floor or the bridge, the ccd video camera array links to each other with image pick-up card by lead, and image pick-up card is installed on computers.
The present invention also has following feature:
1, described ccd video camera array comprises that many model ccd video cameras identical, that have high image resolution connect side by side, and ccd video camera comprises that optical lens is connected with ccd sensor, and optical lens is compatible mutually with ccd sensor; The quantity of ccd video camera is by detected bridge width, detection distance and the decision of ccd video camera imaging rake face size, and wherein a ccd video camera is used for the detection of bridge side, and remaining ccd video camera is used for the detection of bridge bottom vertical.
2, described image pick-up card is a Multiplexing Image Grab Card.
3, the apparent state automatic testing method of a kind of bridge based on the ccd array camera technique, method is as follows:
(1) image filtering denoising method
Bridge bottom apparent image has been brought various noises in transmission course, noise has reduced picture quality to a great extent, and identification brings very big difficulty to defects detection, so before defect image is carried out feature extraction, need carry out smoothing processing to these noises
The disposal route of image smoothing mainly contains averaging of multiple image, median filtering method and weighting neighborhood averaging denoise algorithm, in order to preserve more complete image information, at first adopt the method for grey scale interpolation to carry out interpolation arithmetic to bridge surface imperfection image, then, integrated use median filtering method and weighting neighborhood averaging denoise algorithm are removed the most of noise in the bridge surface imperfection image, and can more effectively preserve good defect image information;
(2) feature extraction matching process
On the basis of filtering and noise reduction method, the extraction of visual defects image characteristic point coupling is by seeking out in contiguous two width of cloth images mutual corresponding relation between the unique point, thereby determine the same unique point coordinate in two width of cloth images respectively, merge so that carry out image mosaic;
The Feature Points Extraction of image commonly used has Harris angle point extraction algorithm, SUSAN angle point extraction algorithm and SIFT feature point extraction algorithm, because the overall performance of SIFT feature point extraction algorithm is better than other algorithm, the unique point of extracting has good geometrical stability, has selected for use SIFT feature point extraction algorithm to extract the unique point of bridge bottom visual defects image.
SIFT feature point extraction algorithm is based on the thought that the characteristics of image yardstick is selected, set up the multiscale space of image, with gaussian kernel function continuous filtering and the down-sampled formation gaussian pyramid image of image by different scale, again two Gaussian image of adjacent yardstick are subtracted each other and obtain difference of Gaussian DOG metric space, each point compares one by one with the point of adjacent yardstick and adjacent position under the DOG metric space, obtain the local extremum position and be residing position of unique point and corresponding yardstick, by surface fitting method unique point is carried out further accurately location, and delete lower point of some contrasts and edge respective point;
The geometric properties that the characteristic matching of defect image is based on apparent image carries out the comparison of similarity, because the sparse property and the uncontinuity of geometric properties itself, the characteristic matching mode can only obtain sparse depth map, needs various interpolating methods could finish the extraction work of view picture depth map at last.On the basis of SIFT feature point extraction algorithm, more various interpolation methods have finally selected for use arest neighbors interpolation method method to carry out characteristic matching, can find the most accurate nearest neighbor distance so adopt the arest neighbors unique point to come that apart from the ratio with inferior neighbour's unique point distance unique point is mated the method for exhaustion.
(3) image mosaic fusion method
The purpose that image mosaic merges is the breakdown diagnosis that the general image that will obtain a seamless bridge bottom visual defects is used for defective, realize that seamless image mosaic need carry out smoothing processing to eliminate the splicing vestige to splicing place, obtains the significantly picture of high-resolution.Realization is in two steps merged in the splicing of bridge bottom visual defects image:
1) image merges
It is that adjacent two width of cloth image mosaics are arrived in the same coordinate space that image merges, and makes two width of cloth images become piece image.After through the characteristics of image coupling, two width of cloth images have produced a pair of set of unique point one to one; In order to splice, need estimate the transformation matrix H between two width of cloth images to two width of cloth images.
Obtaining transformation matrix H need use the back to reflection method with image mapped to be spliced in the reference picture coordinate space; At first 4 boundary coordinates with image to be spliced are mapped in the reference picture coordinate system, thereby determine the map image scope of image to be spliced in reference picture; Then in this scope line by line, scan by pixel, by in being mapped to image coordinate system to be spliced, seeking corresponding input pixel after the inverse matrix of transformation matrix; If the back is a rounded coordinate to the input pixel of mapping, then directly get the output pixel value that this pixel value is the reference picture coordinate system; If the back is not a rounded coordinate to the input pixel of mapping, from 4 pixels that this pixel closes on most, calculate the pixel value of an interpolation as this output pixel point by bilinear interpolation.Finish up to whole picture element scans, obtain the map image of image to be spliced in the reference picture coordinate system.
2) eliminate piece
In the image mosaic process, Non-overlapping Domain belongs to the pixel of image to be spliced and reference picture respectively still can get original pixel grey scale separately, the gray scale of overlapping region interior pixel is chosen then needs careful consideration, for the overlapping region, if choose simply that the pixel of any piece image wherein can noise image fuzzy or make splicing produce the slit, want the removal of images slit, take overlapping region linear transitions method to eliminate the splicing seams problem of overlapping region.
Overlapping region linear transitions method is that hypothesis overlapping region width is L, gets a transition factor sigma may (0≤σ≤1), and the x axle in two width of cloth doubling of the image zones and the maximal value and the minimum value of y axle are divided Xmax, and Xmin and Ymax, Ymin establish the transition factor so
If g
1(x, y), g
2(x, (x y) locates the picture grey scale pixel value at some y) to be respectively reference picture and image to be spliced.(x y) is the image pixel gray-scale value after merging to g.Determine by following formula:
g(x，y)＝σg
1(x，y)+(1-σ)g
2(x，y)
This method makes transition portion smoother, and does not have tangible step;
Calculate the image pixel value of overlapping region by following formula, thereby realize seamlessly transitting of two width of cloth doubling of the image zones, removal of images slit.
Adopt said method that the image of a plurality of ccd video camera collections arranged side by side is spliced the whole photo that fusion can obtain a bridge bottom visual defects image.
(4) defects assessment management method
Several ccd image splicings permeate behind the visual defects image of view picture bridge bottom, use the defect image database that defect image is analyzed, evaluates, managed and safeguards, at first identify the class molded dimension of defective, then associated picture and data are all preserved, be used for the subsequent treatment analysis; The defects assessment management system mainly comprises bridge surface imperfection database and bridge surface imperfection evaluation two parts.
Requirement according to the detection system of bridge surface imperfection, the defect image database is divided into the interpolation defective, deletes defective, checks three functional modules of defective, for to the typical defect image analysis processing, go back the function of defectiveness title, defective generation reason, defect image preservation, data retrieval inquiry, engineering analysis data.
The pre-service of bridge surface imperfection image, discriminance analysis, how much evaluations are mainly carried out in bridge surface imperfection evaluation, and to the information of aspects such as predicting the outcome of following situation; Defect image is being carried out in the Classification and Identification process, selective analysis based on the BP algorithm of " piece " image; Select the training set of the method for artificial generation " piece " image data set for use as neural network; Requirement according to the BP algorithm, respectively to two kinds of BP algorithms based on " piece " image: based on the neural network of image with carry out applied in network performance test based on histogrammic neural network, and determine based on histogrammic neural network algorithm to be the more excellent algorithm of classification of defects identification; Simultaneously, according to the validity feature of bridge surface imperfection, it has been carried out quantizing estimation.
The present invention is based on the apparent condition automatic detection device of bridge and the method thereof of ccd array camera technique, the computer video technology is applied in the bridge surface defects detection, directly provide the differentiation result by reliable CCD automatic data acquisition disposal system, save time, increase work efficiency and precision, and can guarantee the normal operation situation of bridge, significant to the objective evaluation of bridge surface imperfection.
Description of drawings
Fig. 1 is a detection schematic diagram of the present invention;
Fig. 2 is that suspension type of the present invention detects synoptic diagram;
Fig. 3 is that vertical type of the present invention detects synoptic diagram.
Embodiment
The invention will be further described for example below in conjunction with accompanying drawing.
Embodiment 1: the apparent condition automatic detection device of a kind of bridge based on the ccd array camera technique, it is by the ccd video camera array, the scanning frame, image pick-up card, computing machine and inspection vehicle are formed, it is characterized in that: the ccd video camera array is installed on the scanning frame, scanning frame bottom is fixed on the inspection vehicle that is positioned under bridge floor or the bridge, the ccd video camera array links to each other with image pick-up card by lead, image pick-up card is installed on computers, the ccd video camera array is sent to synchronization bridge bottom and side defect image in the computing machine in real time, realizes the synthetic of bridge visual defects image by detecting management software, handle, defect recognition and data management.Described ccd video camera array comprises that many model ccd video cameras identical, that have high image resolution connect side by side, and ccd video camera comprises that optical lens is connected with ccd sensor, and optical lens is compatible mutually with ccd sensor; The quantity of ccd video camera is by detected bridge width, detection distance and the decision of ccd video camera imaging rake face size, and wherein a ccd video camera detects with being used for the bridge side, and the ccd video camera of remainder is used for the bridge bottom vertical and detects.Described image pick-up card is a Multiplexing Image Grab Card.
Embodiment 2: in conjunction with Fig. 1-Fig. 3, the apparent condition automatic detection device of a kind of bridge of the present invention based on the ccd array camera technique, the testing process is as follows: before the inspection, the testing staff is suspended on the scanning frame that carries a plurality of ccd video cameras under the bridge floor on the bridge floor scanning frame bottom being fixed on the inspection vehicle.The ccd video camera array combination by adjusting scanning frame height, makes the image pickup scope of ccd array video camera cover bottom the bridge or the inspection area of sidepiece along with the scanning frame is constantly soaring during detection.The starting point of set checking then, the control detection car traction scanning frame that move on bridge floor drives the ccd array video camera and begins detection.In the testing process, detect evaluation software and show the defect image that the ccd array video camera is taken synchronously by a plurality of windows.The supervisory personnel is by observing display screen, note abnormalities and to suspend the motion of inspection vehicle at any time and carry out the defect characteristic sign, by detecting evaluation automatic defect recognition characteristic type of software and defective locations etc., also the image mosaic that the ccd array video camera is taken can be integrated into piece image according to detecting the needs testing fixture, and print structure surface imperfection image and detection evaluation result, indicate line as image pickup scope, unique point sign and characteristic point information etc.
Embodiment 3:
In conjunction with Fig. 1, using a plurality of ccd array video cameras that the bridge structure visual defects is carried out multiway images when detecting based on the apparent condition automatic detection device of the bridge of ccd array camera technique gathers in real time, because bridge structure visual defects image is subjected to noise serious, pick-up unit combines multiple denoise algorithm defect image has been carried out pre-service; Pick-up unit is cut apart and the feature extraction coupling defective subsequently automatically fast and accurately, and defective edge trend and shape are provided; Then, pick-up unit splices several defect images of ccd array camera acquisition, and melting platform is the general defect image on a bridge polycrystalline substance surface; Pick-up unit carries out statistic of classification in conjunction with practical structures visual defects degree of impairment to structure visual defects damage situation on this basis, and in real time effective decision structure visual defects situation, makes quantitatively and qualitative analysis; At last, pick-up unit carries out signature analysis, discriminator based on the bridge lower surface classification of defects standard of Flame Image Process to defective, has set up the defect management data storehouse of corresponding bridge structure.
The apparent state automatic testing method of a kind of bridge based on the ccd array camera technique is as follows:
1. image filtering denoising method
Bridge bottom apparent image has been brought various noises in transmission course, noise has reduced picture quality to a great extent, identification brings very big difficulty to defects detection, so before defect image is carried out feature extraction, need carry out smoothing processing to these noises.
The disposal route of image smoothing mainly contains averaging of multiple image, median filtering method, neighborhood averaging etc.In order to preserve more complete image information, at first adopt the method for grey scale interpolation to carry out interpolation arithmetic to bridge surface imperfection image.Then, integrated use medium filtering and weighting neighborhood averaging denoise algorithm are removed the most of noise in the bridge surface imperfection image, and can more effectively preserve good defect image information.
2. feature extraction matching process
On the basis of filtering and noise reduction, the extraction of visual defects image characteristic point coupling is by seeking out in contiguous two width of cloth images mutual corresponding relation between the unique point, thereby determine the same unique point coordinate in two width of cloth images respectively, merge so that carry out image mosaic.
The Feature Points Extraction of image commonly used has Harris angle point extraction algorithm, SUSAN angle point extraction algorithm and SIFT feature point extraction algorithm etc., because the overall performance of SIFT operator is better than other operator, the unique point of extracting has good geometrical stability, has selected for use the SIFT algorithm to extract the unique point of bridge bottom visual defects image.
The SIFT algorithm is based on the thought that the characteristics of image yardstick is selected, set up the multiscale space of image, with gaussian kernel function continuous filtering and the down-sampled formation gaussian pyramid image of image by different scale, again two Gaussian image of adjacent yardstick are subtracted each other and obtain difference of Gaussian multiscale space (DOG), each point compares one by one with the point of adjacent yardstick and adjacent position under the DOG metric space, obtain the local extremum position and be residing position of unique point and corresponding yardstick, by surface fitting method unique point is carried out further accurately location, and delete lower point of some contrasts and edge respective point.
The geometric properties that the characteristic matching of defect image is based on apparent image carries out the comparison of similarity, because the sparse property and the uncontinuity of geometric properties itself, the characteristic matching mode can only obtain sparse depth map, needs various interpolating methods could finish the extraction work of view picture depth map at last.On the basis of SIFT feature point extraction algorithm, more various interpolation methods have finally selected for use arest neighbors interpolation method method to carry out characteristic matching, can find the most accurate nearest neighbor distance so adopt the arest neighbors unique point to come that apart from the ratio with inferior neighbour's unique point distance unique point is mated the method for exhaustion, also be the most effective.
3. image mosaic fusion method
The purpose that image mosaic merges is the breakdown diagnosis that the general image that will obtain a seamless bridge bottom visual defects is used for defective, realize that seamless image mosaic need carry out smoothing processing to eliminate the splicing vestige to splicing place, obtains the significantly picture of high-resolution.Realization is in two steps merged in the splicing of bridge bottom visual defects image:
3) image merges
It is that adjacent two width of cloth image mosaics are arrived in the same coordinate space that image merges, and makes two width of cloth images become piece image.After through the characteristics of image coupling, two width of cloth images have produced a pair of set of unique point one to one.In order to splice, need estimate the transformation matrix H between two width of cloth images to two width of cloth images.
Obtaining transformation matrix H need use the back to reflection method with image mapped to be spliced in the reference picture coordinate space.At first 4 boundary coordinates with image to be spliced are mapped in the reference picture coordinate system, thereby determine the map image scope of image to be spliced in reference picture.Then in this scope line by line, scan by pixel, by in being mapped to image coordinate system to be spliced, seeking corresponding input pixel after the inverse matrix of transformation matrix.If the back is a rounded coordinate to the input pixel of mapping, then directly get the output pixel value that this pixel value is the reference picture coordinate system; If the back is not a rounded coordinate to the input pixel of mapping, from 4 pixels that this pixel closes on most, calculate the pixel value of an interpolation as this output pixel point by bilinear interpolation.Finish up to whole picture element scans, obtain the map image of image to be spliced in the reference picture coordinate system.
4) eliminate piece
In the image mosaic process, Non-overlapping Domain belongs to the pixel of image to be spliced and reference picture respectively still can get original pixel grey scale separately, and the gray scale of overlapping region interior pixel is chosen then needs careful consideration.For the overlapping region, if choose simply that the pixel of any piece image wherein can noise image fuzzy or make splicing produce the slit.Want the removal of images slit, take overlapping region linear transitions method to eliminate the splicing seams problem of overlapping region.
Overlapping region linear transitions method is that hypothesis overlapping region width is L, gets a transition factor sigma may (0≤σ≤1), and the x axle in two width of cloth doubling of the image zones and the maximal value and the minimum value of y axle are divided Xmax, Xmin.And Ymax, Ymin.Establish the transition factor so
If g
1(x, y), g
2(x, (x y) locates the picture grey scale pixel value at some y) to be respectively reference picture and image to be spliced.(x y) is the image pixel gray-scale value after merging to g.Determine by following formula:
g(x，y)＝σg
1(x，y)+(1-σ)g
2(x，y)
This method makes transition portion smoother, and does not have tangible step.
Calculate the image pixel value of overlapping region by following formula, thereby realize seamlessly transitting of two width of cloth doubling of the image zones, removal of images slit.
Adopt said method that the image of a plurality of ccd video camera collections arranged side by side is spliced the whole photo that fusion can obtain a bridge bottom visual defects image.
4. defects assessment management method
Several ccd image splicings permeate behind the visual defects image of view picture bridge bottom, use the defect image database that defect image is analyzed, evaluates, managed and safeguards, at first identify the class molded dimension of defective, then associated picture and data are all preserved, be used for the subsequent treatment analysis.The defects assessment management system mainly comprises bridge surface imperfection database and bridge surface imperfection evaluation two parts.
Requirement according to the detection system of bridge surface imperfection, the defect image database is divided into the interpolation defective, deletes defective, checks three functional modules of defective, for to the typical defect image analysis processing, go back functions such as defectiveness title, defective generation reason, defect image preservation, data retrieval inquiry, engineering analysis data.
The pre-service of bridge surface imperfection image, discriminance analysis, how much evaluations are mainly carried out in bridge surface imperfection evaluation, and to the information of aspects such as predicting the outcome of following situation.Defect image is being carried out in the Classification and Identification process, selective analysis based on the BP algorithm of " piece " image.Select the training set of the method for artificial generation " piece " image data set for use as neural network.Requirement according to the BP algorithm, respectively to two kinds of BP algorithms based on " piece " image: based on the neural network of image with carry out applied in network performance test based on histogrammic neural network, and determine based on histogrammic neural network algorithm to be the more excellent algorithm of classification of defects identification.Simultaneously, according to the validity feature of bridge surface imperfection, it has been carried out quantizing estimation.
Claims (4)
1. apparent condition automatic detection device of the bridge based on the ccd array camera technique, it is by the ccd video camera array, the scanning frame, image pick-up card, computing machine, detection management software and inspection vehicle are formed, it is characterized in that: the ccd video camera array is installed on the scanning frame, scanning frame bottom is fixed on the inspection vehicle that is positioned under bridge floor or the bridge, the ccd video camera array links to each other with image pick-up card by lead, image pick-up card is installed on computers, the ccd video camera array is sent to synchronization bridge bottom and side defect image in the computing machine in real time, realizes the synthetic of bridge visual defects image by detecting management software, handle, defect recognition and data management.
2. the apparent condition automatic detection device of a kind of bridge according to claim 1 based on the ccd array camera technique, it is characterized in that: described ccd video camera array comprises that many model ccd video cameras identical, that have high image resolution connect side by side, ccd video camera comprises that optical lens is connected with ccd sensor, and optical lens is compatible mutually with ccd sensor; The quantity of ccd video camera is by detected bridge width, detection distance and the decision of ccd video camera imaging rake face size, and wherein a ccd video camera is used for the detection of bridge side, and remaining ccd video camera is used for the detection of bridge bottom vertical;
3. the apparent condition automatic detection device of a kind of bridge according to claim 1 based on the ccd array camera technique, it is characterized in that: described image pick-up card is a Multiplexing Image Grab Card;
4. apparent state automatic testing method of the bridge based on the ccd array camera technique is characterized in that method is as follows:
(1) image filtering denoising method
Bridge bottom apparent image has been brought various noises in transmission course, noise has reduced picture quality to a great extent, and identification brings very big difficulty to defects detection, so before defect image is carried out feature extraction, need carry out smoothing processing to these noises
The disposal route of image smoothing mainly contains averaging of multiple image, median filtering method and weighting neighborhood averaging denoise algorithm, in order to preserve more complete image information, at first adopt the method for grey scale interpolation to carry out interpolation arithmetic to bridge surface imperfection image, then, integrated use median filtering method and weighting neighborhood averaging denoise algorithm are removed the most of noise in the bridge surface imperfection image, and can more effectively preserve good defect image information;
(2) feature extraction matching process
On the basis of filtering and noise reduction method, the extraction of visual defects image characteristic point coupling is by seeking out in contiguous two width of cloth images mutual corresponding relation between the unique point, thereby determine the same unique point coordinate in two width of cloth images respectively, merge so that carry out image mosaic;
The Feature Points Extraction of image commonly used has Harris angle point extraction algorithm, SUSAN angle point extraction algorithm and SIFT feature point extraction algorithm, because the overall performance of SIFT feature point extraction algorithm is better than other algorithm, the unique point of extracting has good geometrical stability, has selected for use SIFT feature point extraction algorithm to extract the unique point of bridge bottom visual defects image;
SIFT feature point extraction algorithm is based on the thought that the characteristics of image yardstick is selected, set up the multiscale space of image, with gaussian kernel function continuous filtering and the down-sampled formation gaussian pyramid image of image by different scale, again two Gaussian image of adjacent yardstick are subtracted each other and obtain difference of Gaussian DOG metric space, each point compares one by one with the point of adjacent yardstick and adjacent position under the DOG metric space, obtain the local extremum position and be residing position of unique point and corresponding yardstick, by surface fitting method unique point is carried out further accurately location, and delete lower point of some contrasts and edge respective point;
The geometric properties that the characteristic matching of defect image is based on apparent image carries out the comparison of similarity, because the sparse property and the uncontinuity of geometric properties itself, the characteristic matching mode can only obtain sparse depth map, needs various interpolating methods could finish the extraction work of view picture depth map at last; On the basis of SIFT feature point extraction algorithm, more various interpolation methods have finally selected for use arest neighbors interpolation method method to carry out characteristic matching, can find the most accurate nearest neighbor distance so adopt the arest neighbors unique point to come that apart from the ratio with inferior neighbour's unique point distance unique point is mated the method for exhaustion;
(3) image mosaic fusion method
The purpose that image mosaic merges is the breakdown diagnosis that the general image that will obtain a seamless bridge bottom visual defects is used for defective, realize that seamless image mosaic need carry out smoothing processing to eliminate the splicing vestige to splicing place, obtains the significantly picture of high-resolution; Realization is in two steps merged in the splicing of bridge bottom visual defects image:
1) image merges
It is that adjacent two width of cloth image mosaics are arrived in the same coordinate space that image merges, and makes two width of cloth images become piece image; After through the characteristics of image coupling, two width of cloth images have produced a pair of set of unique point one to one; In order to splice, need estimate the transformation matrix H between two width of cloth images to two width of cloth images;
Obtaining transformation matrix H need use the back to reflection method with image mapped to be spliced in the reference picture coordinate space; At first 4 boundary coordinates with image to be spliced are mapped in the reference picture coordinate system, thereby determine the map image scope of image to be spliced in reference picture; Then in this scope line by line, scan by pixel, by in being mapped to image coordinate system to be spliced, seeking corresponding input pixel after the inverse matrix of transformation matrix; If the back is a rounded coordinate to the input pixel of mapping, then directly get the output pixel value that this pixel value is the reference picture coordinate system; If the back is not a rounded coordinate to the input pixel of mapping, from 4 pixels that this pixel closes on most, calculate the pixel value of an interpolation as this output pixel point by bilinear interpolation; Finish up to whole picture element scans, obtain the map image of image to be spliced in the reference picture coordinate system;
2) eliminate piece
In the image mosaic process, Non-overlapping Domain belongs to the pixel of image to be spliced and reference picture respectively still can get original pixel grey scale separately, the gray scale of overlapping region interior pixel is chosen then needs careful consideration, for the overlapping region, if choose simply that the pixel of any piece image wherein can noise image fuzzy or make splicing produce the slit, want the removal of images slit, take overlapping region linear transitions method to eliminate the splicing seams problem of overlapping region;
Overlapping region linear transitions method is that hypothesis overlapping region width is L, gets a transition factor sigma may (0≤σ≤1), and the x axle in two width of cloth doubling of the image zones and the maximal value and the minimum value of y axle are divided Xmax, and Xmin and Ymax, Ymin establish the transition factor so
If g
1(x, y), g
2(x, (x y) locates the picture grey scale pixel value at some y) to be respectively reference picture and image to be spliced;
(x y) is the image pixel gray-scale value after merging to g; Determine by following formula:
g(x，y)＝σg
1(x，y)+(1-σ)g
2(x，y)
This method makes transition portion smoother, and does not have tangible step;
Calculate the image pixel value of overlapping region by following formula, thereby realize seamlessly transitting of two width of cloth doubling of the image zones, removal of images slit;
Adopt said method that the image of a plurality of ccd video camera collections arranged side by side is spliced the whole photo that fusion can obtain a bridge bottom visual defects image;
(4) defects assessment management method
Several ccd image splicings permeate behind the visual defects image of view picture bridge bottom, use the defect image database that defect image is analyzed, evaluates, managed and safeguards, at first identify the class molded dimension of defective, then associated picture and data are all preserved, be used for the subsequent treatment analysis; The defects assessment management system mainly comprises bridge surface imperfection database and bridge surface imperfection evaluation two parts;
Requirement according to the detection system of bridge surface imperfection, the defect image database is divided into the interpolation defective, deletes defective, checks three functional modules of defective, for to the typical defect image analysis processing, go back the function of defectiveness title, defective generation reason, defect image preservation, data retrieval inquiry, engineering analysis data;
The pre-service of bridge surface imperfection image, discriminance analysis, how much evaluations are mainly carried out in bridge surface imperfection evaluation, and to the information of aspects such as predicting the outcome of following situation; Defect image is being carried out in the Classification and Identification process, selective analysis based on the BP algorithm of " piece " image; Select the training set of the method for artificial generation " piece " image data set for use as neural network; Requirement according to the BP algorithm, respectively to two kinds of BP algorithms based on " piece " image: based on the neural network of image with carry out applied in network performance test based on histogrammic neural network, and determine based on histogrammic neural network algorithm to be the more excellent algorithm of classification of defects identification; Simultaneously, according to the validity feature of bridge surface imperfection, it has been carried out quantizing estimation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100230283A CN102183525A (en) | 2011-01-20 | 2011-01-20 | Bridge apparent state automatic detection device and method based on charge coupled device (CCD) array photographic technique |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100230283A CN102183525A (en) | 2011-01-20 | 2011-01-20 | Bridge apparent state automatic detection device and method based on charge coupled device (CCD) array photographic technique |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102183525A true CN102183525A (en) | 2011-09-14 |
Family
ID=44569746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011100230283A Pending CN102183525A (en) | 2011-01-20 | 2011-01-20 | Bridge apparent state automatic detection device and method based on charge coupled device (CCD) array photographic technique |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102183525A (en) |
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102854194A (en) * | 2012-09-11 | 2013-01-02 | 中南大学 | Object surface defect detection method and apparatus based on linear array CCD |
| CN103034207A (en) * | 2012-12-14 | 2013-04-10 | 江苏飞尚安全监测咨询有限公司 | Infrastructure health monitoring system and implementation process thereof |
| CN103095990A (en) * | 2012-12-18 | 2013-05-08 | 武汉烽火众智数字技术有限责任公司 | Ultra high definition (UHD) video camera provided with distributed image capture units |
| CN104486613A (en) * | 2014-11-19 | 2015-04-01 | 江苏影速光电技术有限公司 | Method for solving abnormity of data transmission of image acquisition system in movement process |
| CN104931144A (en) * | 2015-02-06 | 2015-09-23 | 东南大学 | Coal-fired power plant coal conveyor belt temperature remote monitoring system |
| CN104990935A (en) * | 2015-07-27 | 2015-10-21 | 南阳理工学院 | Bridge and building health condition monitoring device |
| CN105472330A (en) * | 2015-12-01 | 2016-04-06 | 芜湖纯元光电设备技术有限公司 | Method for solving problem of transmission abnormity of image collection data in laser transmission |
| CN106034196A (en) * | 2015-03-10 | 2016-10-19 | 青岛通产软件科技有限公司 | Multi-visual-angle image integration acquisition system |
| CN106033541A (en) * | 2015-03-10 | 2016-10-19 | 青岛通产软件科技有限公司 | Electronic number rubbing device and method for motor vehicle |
| CN106447696A (en) * | 2016-09-29 | 2017-02-22 | 郑州轻工业学院 | Bidirectional SIFT (scale invariant feature transformation) flow motion evaluation-based large-displacement target sparse tracking method |
| CN106530274A (en) * | 2016-10-11 | 2017-03-22 | 昆明理工大学 | Steel bridge crack positioning method |
| CN106575435A (en) * | 2014-05-26 | 2017-04-19 | 赛峰飞机发动机公司 | Method and device for estimation of quality index of 3-d image of piece of composite material |
| CN106918598A (en) * | 2017-03-08 | 2017-07-04 | 河海大学 | Bridge pavement strain and crack detection analysis system and method based on digital picture |
| CN106993157A (en) * | 2017-04-05 | 2017-07-28 | 宇龙计算机通信科技(深圳)有限公司 | A kind of intelligent control method and device based on dual camera |
| CN107589122A (en) * | 2017-09-22 | 2018-01-16 | 成都市鹰诺实业有限公司 | Optical scanner image detection device and detection method |
| CN107941816A (en) * | 2017-12-29 | 2018-04-20 | 苏州德创测控科技有限公司 | Portable appearance delection device and appearance detecting method |
| CN108133070A (en) * | 2017-09-19 | 2018-06-08 | 广州市建筑科学研究院有限公司 | A kind of appraisal procedure and system of the bridge health situation based on radial basis function neural network |
| CN108140218A (en) * | 2015-09-10 | 2018-06-08 | 富士胶片株式会社 | Integrity decision maker, integrity determination method and integrity decision procedure |
| CN108459030A (en) * | 2018-02-08 | 2018-08-28 | 东华大学 | One kind being applied to non-planar plastic smooth surface flaw on-line measuring device and method |
| CN108921848A (en) * | 2018-09-29 | 2018-11-30 | 长安大学 | Bridge Defect Detecting device and detection image joining method based on more mesh cameras |
| CN108956638A (en) * | 2018-04-27 | 2018-12-07 | 湖南文理学院 | A kind of evaluation detection system for civil engineering structure visual defects |
| CN109164112A (en) * | 2018-09-26 | 2019-01-08 | 深圳森阳环保材料科技有限公司 | A kind of cable surface defects detection system based on unmanned plane |
| CN109239076A (en) * | 2018-08-29 | 2019-01-18 | 西安理工大学 | A kind of sewing thread trace defect inspection method based on machine vision |
| CN109313163A (en) * | 2016-11-29 | 2019-02-05 | 瓦锡兰芬兰有限公司 | It is controlled using the ultrasonic wave mass of filtered image data |
| CN109668547A (en) * | 2019-01-30 | 2019-04-23 | 湖北辉创重型工程有限公司 | A kind of bridge intelligence inspection system |
| CN110084844A (en) * | 2019-04-25 | 2019-08-02 | 中国民航大学 | A kind of airfield pavement crack detection method based on depth camera |
| CN110147781A (en) * | 2019-05-29 | 2019-08-20 | 重庆交通大学 | Bridge vibration mode based on machine learning visualizes damnification recognition method |
| CN110596116A (en) * | 2019-07-23 | 2019-12-20 | 浙江科技学院 | Vehicle surface flaw detection method and system |
| CN110672620A (en) * | 2019-10-08 | 2020-01-10 | 英特尔产品（成都）有限公司 | Chip defect detection method and system |
| CN112164051A (en) * | 2020-09-29 | 2021-01-01 | 中国船舶重工集团公司第七二四研究所 | Radar antenna area array liquid leakage detection device and method based on image analysis |
| CN115086569A (en) * | 2022-06-10 | 2022-09-20 | 湖南康桥智能科技有限公司 | Method for acquiring images of bottom of super-large bridge based on networking camera |
| CN116597125A (en) * | 2023-05-24 | 2023-08-15 | 中国公路工程咨询集团有限公司 | Bridge splicing remote sensing control system and method based on image recognition |
| CN117351011A (en) * | 2023-12-04 | 2024-01-05 | 歌尔股份有限公司 | Screen defect detection method, apparatus, and readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004012454A (en) * | 2002-06-11 | 2004-01-15 | Toyo High Mech Kk | Method and device for detecting exfoliation inside concrete structure |
| CN201473886U (en) * | 2009-08-26 | 2010-05-19 | 江苏省交通科学研究院股份有限公司 | Bridge inspection vehicle |
| CN101713167A (en) * | 2009-10-23 | 2010-05-26 | 周劲宇 | Bridge structural health monitoring car |
-
2011
- 2011-01-20 CN CN2011100230283A patent/CN102183525A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004012454A (en) * | 2002-06-11 | 2004-01-15 | Toyo High Mech Kk | Method and device for detecting exfoliation inside concrete structure |
| CN201473886U (en) * | 2009-08-26 | 2010-05-19 | 江苏省交通科学研究院股份有限公司 | Bridge inspection vehicle |
| CN101713167A (en) * | 2009-10-23 | 2010-05-26 | 周劲宇 | Bridge structural health monitoring car |
Non-Patent Citations (3)
| Title |
|---|
| 刘恩锴: "基于特征的图像拼接技术研究", 《中国优秀硕士学位论文全文数据库》 * |
| 曾燕华: "基于计算机视觉的桥梁表面缺陷检测技术研究", 《中国优秀硕士学位论文全文数据库》 * |
| 顾费勇: "基于图像的自适应图像拼接算法研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (43)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102854194A (en) * | 2012-09-11 | 2013-01-02 | 中南大学 | Object surface defect detection method and apparatus based on linear array CCD |
| CN102854194B (en) * | 2012-09-11 | 2014-11-05 | 中南大学 | Object surface defect detection method and apparatus based on linear array CCD |
| CN103034207A (en) * | 2012-12-14 | 2013-04-10 | 江苏飞尚安全监测咨询有限公司 | Infrastructure health monitoring system and implementation process thereof |
| CN103095990A (en) * | 2012-12-18 | 2013-05-08 | 武汉烽火众智数字技术有限责任公司 | Ultra high definition (UHD) video camera provided with distributed image capture units |
| US10402678B2 (en) | 2014-05-26 | 2019-09-03 | Safran Aircraft Engines | Method and a device for estimating a quality index for a 3D image of a composite material part |
| CN106575435A (en) * | 2014-05-26 | 2017-04-19 | 赛峰飞机发动机公司 | Method and device for estimation of quality index of 3-d image of piece of composite material |
| CN104486613A (en) * | 2014-11-19 | 2015-04-01 | 江苏影速光电技术有限公司 | Method for solving abnormity of data transmission of image acquisition system in movement process |
| CN104931144A (en) * | 2015-02-06 | 2015-09-23 | 东南大学 | Coal-fired power plant coal conveyor belt temperature remote monitoring system |
| CN106033541A (en) * | 2015-03-10 | 2016-10-19 | 青岛通产软件科技有限公司 | Electronic number rubbing device and method for motor vehicle |
| CN106034196A (en) * | 2015-03-10 | 2016-10-19 | 青岛通产软件科技有限公司 | Multi-visual-angle image integration acquisition system |
| CN104990935A (en) * | 2015-07-27 | 2015-10-21 | 南阳理工学院 | Bridge and building health condition monitoring device |
| CN108140218A (en) * | 2015-09-10 | 2018-06-08 | 富士胶片株式会社 | Integrity decision maker, integrity determination method and integrity decision procedure |
| CN108140218B (en) * | 2015-09-10 | 2022-03-25 | 富士胶片株式会社 | Health degree determination device, health degree determination method, and health degree determination program |
| CN105472330A (en) * | 2015-12-01 | 2016-04-06 | 芜湖纯元光电设备技术有限公司 | Method for solving problem of transmission abnormity of image collection data in laser transmission |
| CN106447696A (en) * | 2016-09-29 | 2017-02-22 | 郑州轻工业学院 | Bidirectional SIFT (scale invariant feature transformation) flow motion evaluation-based large-displacement target sparse tracking method |
| CN106530274A (en) * | 2016-10-11 | 2017-03-22 | 昆明理工大学 | Steel bridge crack positioning method |
| CN106530274B (en) * | 2016-10-11 | 2019-04-12 | 昆明理工大学 | A kind of localization method of girder steel crackle |
| CN109313163A (en) * | 2016-11-29 | 2019-02-05 | 瓦锡兰芬兰有限公司 | It is controlled using the ultrasonic wave mass of filtered image data |
| CN106918598A (en) * | 2017-03-08 | 2017-07-04 | 河海大学 | Bridge pavement strain and crack detection analysis system and method based on digital picture |
| CN106918598B (en) * | 2017-03-08 | 2019-06-21 | 河海大学 | Bridge pavement strain and crack detection analysis system and method based on digital picture |
| CN106993157A (en) * | 2017-04-05 | 2017-07-28 | 宇龙计算机通信科技(深圳)有限公司 | A kind of intelligent control method and device based on dual camera |
| CN108133070A (en) * | 2017-09-19 | 2018-06-08 | 广州市建筑科学研究院有限公司 | A kind of appraisal procedure and system of the bridge health situation based on radial basis function neural network |
| CN107589122A (en) * | 2017-09-22 | 2018-01-16 | 成都市鹰诺实业有限公司 | Optical scanner image detection device and detection method |
| CN107941816A (en) * | 2017-12-29 | 2018-04-20 | 苏州德创测控科技有限公司 | Portable appearance delection device and appearance detecting method |
| CN108459030A (en) * | 2018-02-08 | 2018-08-28 | 东华大学 | One kind being applied to non-planar plastic smooth surface flaw on-line measuring device and method |
| CN108956638A (en) * | 2018-04-27 | 2018-12-07 | 湖南文理学院 | A kind of evaluation detection system for civil engineering structure visual defects |
| CN109239076A (en) * | 2018-08-29 | 2019-01-18 | 西安理工大学 | A kind of sewing thread trace defect inspection method based on machine vision |
| CN109164112A (en) * | 2018-09-26 | 2019-01-08 | 深圳森阳环保材料科技有限公司 | A kind of cable surface defects detection system based on unmanned plane |
| CN108921848A (en) * | 2018-09-29 | 2018-11-30 | 长安大学 | Bridge Defect Detecting device and detection image joining method based on more mesh cameras |
| CN109668547A (en) * | 2019-01-30 | 2019-04-23 | 湖北辉创重型工程有限公司 | A kind of bridge intelligence inspection system |
| CN110084844A (en) * | 2019-04-25 | 2019-08-02 | 中国民航大学 | A kind of airfield pavement crack detection method based on depth camera |
| CN110084844B (en) * | 2019-04-25 | 2023-03-28 | 中国民航大学 | Airport pavement crack detection method based on depth camera |
| CN110147781A (en) * | 2019-05-29 | 2019-08-20 | 重庆交通大学 | Bridge vibration mode based on machine learning visualizes damnification recognition method |
| CN110596116A (en) * | 2019-07-23 | 2019-12-20 | 浙江科技学院 | Vehicle surface flaw detection method and system |
| CN110672620A (en) * | 2019-10-08 | 2020-01-10 | 英特尔产品（成都）有限公司 | Chip defect detection method and system |
| CN110672620B (en) * | 2019-10-08 | 2022-08-26 | 英特尔产品（成都）有限公司 | Chip defect detection method and system |
| CN112164051A (en) * | 2020-09-29 | 2021-01-01 | 中国船舶重工集团公司第七二四研究所 | Radar antenna area array liquid leakage detection device and method based on image analysis |
| CN115086569A (en) * | 2022-06-10 | 2022-09-20 | 湖南康桥智能科技有限公司 | Method for acquiring images of bottom of super-large bridge based on networking camera |
| CN115086569B (en) * | 2022-06-10 | 2024-04-19 | 湖南康桥智能科技有限公司 | Extra-large bridge bottom image acquisition method based on networking camera |
| CN116597125A (en) * | 2023-05-24 | 2023-08-15 | 中国公路工程咨询集团有限公司 | Bridge splicing remote sensing control system and method based on image recognition |
| CN116597125B (en) * | 2023-05-24 | 2023-11-21 | 中国公路工程咨询集团有限公司 | Bridge splicing remote sensing control system and method based on image recognition |
| CN117351011A (en) * | 2023-12-04 | 2024-01-05 | 歌尔股份有限公司 | Screen defect detection method, apparatus, and readable storage medium |
| CN117351011B (en) * | 2023-12-04 | 2024-03-12 | 歌尔股份有限公司 | Screen defect detection method, apparatus, and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102183525A (en) | Bridge apparent state automatic detection device and method based on charge coupled device (CCD) array photographic technique | |
| US11221107B2 (en) | Method for leakage detection of underground pipeline corridor based on dynamic infrared thermal image processing | |
| Song et al. | Real-time tunnel crack analysis system via deep learning | |
| Li et al. | Integrated processing of image and GPR data for automated pothole detection | |
| Tang et al. | Automatic crack detection and segmentation using a hybrid algorithm for road distress analysis | |
| JP6620477B2 (en) | Method and program for detecting cracks in concrete | |
| Li et al. | Long-distance precision inspection method for bridge cracks with image processing | |
| CN102175700B (en) | Method for detecting welding seam segmentation and defects of digital X-ray images | |
| CN104112269B (en) | A kind of solar battery laser groove parameter detection method and system based on machine vision | |
| CN109489724B (en) | Tunnel train safe operation environment comprehensive detection device and detection method | |
| CN105160652A (en) | Handset casing testing apparatus and method based on computer vision | |
| CN113781537B (en) | Rail elastic strip fastener defect identification method and device and computer equipment | |
| CN110390256B (en) | Asphalt pavement crack extraction method | |
| CN112858321A (en) | Steel plate surface defect detection system and method based on linear array CCD | |
| Lee et al. | Diagnosis of crack damage on structures based on image processing techniques and R-CNN using unmanned aerial vehicle (UAV) | |
| CN204039886U (en) | A kind of pavement damage crack detection system based on multiple stage camera stereoscopic shooting | |
| Fu et al. | Research on image-based detection and recognition technologies for cracks on rail surface | |
| WO2021227287A1 (en) | Method and apparatus for honeycomb product regularity inspection on production line | |
| Medina et al. | Automated visual inspection of road surface cracks | |
| CN115841477A (en) | Multi-mode visual fusion ballastless track structure hidden defect detection method and system | |
| CN117422699A (en) | Highway detection method, highway detection device, computer equipment and storage medium | |
| Kong et al. | Detecting type and size of road crack with the smartphone | |
| CN116912805A (en) | Well lid abnormity intelligent detection and identification method and system based on unmanned sweeping vehicle | |
| CN104992426B (en) | A kind of multi-layer image blending algorithm for light field micro-imaging | |
| Zhang et al. | Recognition of pavement surface crack |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110914 |