CN110487497B - Bridge crack identification method based on recursive search - Google Patents
Bridge crack identification method based on recursive search Download PDFInfo
- Publication number
- CN110487497B CN110487497B CN201910673031.6A CN201910673031A CN110487497B CN 110487497 B CN110487497 B CN 110487497B CN 201910673031 A CN201910673031 A CN 201910673031A CN 110487497 B CN110487497 B CN 110487497B
- Authority
- CN
- China
- Prior art keywords
- crack
- segment
- fracture
- image
- fragment
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 239000012634 fragment Substances 0.000 claims abstract description 110
- 230000011218 segmentation Effects 0.000 claims abstract description 38
- 238000004140 cleaning Methods 0.000 claims abstract description 7
- 238000012216 screening Methods 0.000 claims abstract description 7
- 239000013598 vector Substances 0.000 claims description 30
- 230000009191 jumping Effects 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 15
- 238000001914 filtration Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000002474 experimental method Methods 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 9
- 238000003709 image segmentation Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000005286 illumination Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M5/00—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings
- G01M5/0008—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings of bridges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M5/00—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings
- G01M5/0033—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings by determining damage, crack or wear
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Geometry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a bridge crack identification method based on recursive search, which comprises the following steps: converting a crack image to be identified into a crack gray image G, and performing threshold segmentation on the crack image by adopting a local dynamic threshold segmentation method to obtain a segmentation image T; performing feature screening on the segmented image T to obtain a suspicious crack segment set; initializing parameters of the suspicious fracture fragment set one by one; according to the longitudinal coordinate value of the rear end point of the suspicious fracture fragment set, reordering the suspicious fracture fragment set according to a descending order to obtain a fracture fragment set P which is subjected to ordering; searching and identifying the crack segment set P through a crack identification algorithm based on recursive search, and outputting detected crack data which completes the recursion; and cleaning the detected crack data and outputting the data as a recognition result. The method can effectively eliminate the influence of the interference characteristics on crack identification.
Description
Technical Field
The invention relates to the technical field of computer vision, in particular to a bridge crack identification method based on recursive search.
Background
The bridge crack identification technology is an engineering discipline which shoots an object through a camera, obtains image information and then carries out image analysis so as to obtain information such as the pose, the size and the like of the object. For the measurement object, adopt computer identification and processing can not receive the influence of artificial factor, the precision is not restricted by reference object precision such as measurement scale, and compare in traditional artifical scale measurement, has advantages such as quick, accurate, non-contact and low cost, can greatly reduce the cost of labor, effectively improves production efficiency.
When bridge cracks are identified, a detection target is generally adopted, and then non-contact measurement is carried out on the target. When the computer detects the target of the bridge crack image, an image segmentation processing method is often used, and automatic threshold segmentation methods such as a fixed threshold segmentation method, an edge segmentation method, a maximum inter-class variance and the like can be adopted in the image segmentation processing process. However, the fixed threshold segmentation method has a problem of being seriously influenced by illumination change, the edge segmentation method has a problem of being seriously influenced by object texture, and the maximum inter-class variance method can effectively inhibit the influence of illumination change, but when the smooth surface of a target object part generates mirror reflection or the gray levels of the object surface and the background are in a state of 3 or more than 3 obvious gray levels, only a region with reflection or the highest bright gray level is easily segmented, and other parts of the object are ignored. In addition, when the target shape of the crack image is elongated, image segmentation may cause fracture of the target crack, so that the entire crack appears as a series of crack segments, and when the interference features in the image are equivalent to the crack segments in size, form, or color, it is difficult to effectively detect the crack target.
Disclosure of Invention
The invention provides a bridge crack identification method based on recursive search, aiming at overcoming the defect that the influence of interference characteristics on crack identification cannot be eliminated in the prior art.
In order to solve the technical problems, the technical scheme of the invention is as follows:
a bridge crack identification method based on recursive search comprises the following steps:
s1: converting a crack image to be identified into a crack gray image G, and performing threshold segmentation on the crack image by adopting a local dynamic threshold segmentation method to obtain a segmentation image T;
s2: performing characteristic screening on the segmentation image T to obtain a suspicious crack segment set Cp={CpiH, wherein i is the sequence number of the suspicious crack segment;
s3: collecting C the suspicious fracture fragmentspInitializing parameters one by one;
s4: collecting C according to the suspicious crack segmentspThe longitudinal coordinate value of the rear end point, and the suspicious fracture segment set CpReordering according to descending order to obtain a result of completed orderingA fracture fragment set P;
s5: searching and identifying the crack segment set P through a crack identification algorithm based on recursive search, and outputting detected crack data which completes the recursion;
s6: and cleaning the detected crack data and outputting the data as a recognition result.
According to the technical scheme, firstly, a local dynamic threshold segmentation method is adopted to carry out threshold segmentation on a crack image, suspicious crack segments are screened out, then a matching threshold is set, the robustness of the method is effectively improved, finally, interference characteristics which are equivalent to the size, the shape or the color of the crack segments are discharged through a crack identification algorithm based on recursive search, and a plurality of cracks are effectively searched, detected and combined.
Preferably, in the step S1, the specific step of performing threshold segmentation on the crack image by using a local dynamic threshold segmentation method is as follows:
s11: carrying out mean value filtering on the crack gray level image G to obtain a mean value filtering image M;
s12: and (3) making the image T be | M-G |, traversing the image T, and when the gray value of a pixel point of the image T is greater than a preset threshold offset, setting the gray value of the pixel point to be 255 to obtain a segmented image T which is subjected to threshold segmentation.
Preferably, in the step S2, the specific steps of feature screening for the segmented image T are as follows:
s21: detecting connected component contour C of segmented image TkWherein k represents the serial number of the detected connected domain profile;
s22: calculating the connected component contour CkThe length-width ratio f and the circularity e of the minimum circumscribed rectangle are calculated, whether suspicious crack judgment conditions are met is judged, and if yes, the connected domain contour C is subjected to classificationkMarking as suspicious fracture fragment CpiWherein the aspect ratio f and the circularity e are calculated as follows:
s23: after traversing all connected domains, marking the connected domains as suspicious crack segments CpiIntegration into a set of suspicious fracture fragments Cp={Cpi}。
Preferably, the suspicious crack determination condition is that the connected domain profile CkAspect ratio f > t0Or connected domain profile CkThe circularity e of (a) satisfies (t)1E and f > t2Wherein, t0、t1、t2Is a predetermined threshold value obtained by a plurality of experiments, and t0>t2。
Preferably, in the step S3, the initialization parameters include a fracture fragment CpiThe end point coordinates, the direction vectors, the lengths, the mark positions, the segment distances and the segment included angles of the suspicious crack segments, wherein the suspicious crack segments are collectedpThe specific steps of initializing the parameters one by one are as follows:
definition of fracture fragment CpiThe front end point of (2) is the point with the minimum longitudinal coordinate value in the contour;
definition of fracture fragment CpiThe rear end point of (2) is the point with the maximum longitudinal coordinate value in the profile;
definition of fracture fragment CpiThe direction vector of (1) is the long side direction vector of the minimum circumscribed rectangle, and the direction of the direction vector is from one end where the rear endpoint is located to one end where the front endpoint is located;
definition of fracture fragment CpiLength L (C) ofpi) The length of the long side of the minimum circumscribed rectangle is the length of the long side;
definition of fracture fragment CpiThe fracture flag of (1);
definition of fracture fragment CpiAnd fracture fragment CpjFragment distance D (C) ofpi,Cpj) Is a fracture fragment CpiFront end point and fracture fragment C ofpjThe Euclidean distance between the rear end points of (a), wherein j is the sequence number of the crack segment;
definition of fracture fragment CpiAnd fracture fragment CpjAngle of segment A (C)pi,Cpj) Is a vectorAnd fracture fragment CpjAngle of direction vector of (1), wherein the vectorTo be a fracture fragment CpiStarting from the leading end of (A), and a fracture fragment CpjThe rear end points of (a) are concatenated into a vector.
Preferably, in the step S5, the specific steps of performing search identification on the fracture fragment set P by the fracture identification algorithm based on recursive search are as follows:
s51: initializing the number m of cracks as 1;
s52: taking a crack segment with the first crack flag bit flag being-1 from the crack segment set P as a currently processed segment P0Setting the crack flag bit as m, and judging the current processing segment P except the current processing segment P in the crack segment set P0If the fracture flag of other fracture fragments is equal to-1, executing the step S53; if not, jumping to execute the step S58;
s53: taking a crack segment with a second crack flag position of flag-1 from the crack segment set P as a segment P to be matchedn;
S54: calculating the currently processed segment P0And a segment P to be matchednFragment distance D (P) of0,Pn) And its length L (P)0)、L(Pn) Judging the segment distance D (P)0,Pn) Whether the distance is smaller than a preset distance threshold doff or not, if yes, executing a step S55; if not, further judging the segment distance D (P)0,Pn) Whether or not D (P) is satisfied0,Pn)<5×min(L(P0),L(Pn) If yes, jumping to the step of S56; if not, making m equal to m +1, and then jumping to execute the step S52;
s55: calculating the currently processed segment P0And a segment P to be matchednAngle of segment A (P)0,Pn) Judging the included angle A (P)0,Pn) Whether or not it is less than a preset angle threshold aoffIf yes, the segment P to be matched is determinednThe crack flag bit is set as m, and the current segment P to be matched is set asnAs the currently processed segment P0=PnTaking the next crack segment as the segment P to be matchedn=Pn+1And jumping to execute the step of S57; if not, go to step S56;
s56: judging the current segment P to be matchednIf the element is the last element in the fracture fragment set P, if so, making m equal to m +1, and then jumping to execute the step S52; if not, taking the next crack segment as the segment P to be matchedn=Pn+1And executing the step of S57;
s57: judging the current segment P to be matchednIf the crack segment set P is not in the crack segment set P, jumping to execute the step S54, and if not, executing the step S58;
s58: judging whether the number m of the cracks is equal to the number of elements in the crack segment set P, if so, outputting the crack segment set P after the recurrence as recognized crack data; if not, combining the fracture fragments with the same fracture flag into a new fracture fragment, combining the corresponding fracture fragments into a new combined element in the set P, initializing the parameters of the new combined element, and then skipping to execute S51 to perform recursive search again.
Preferably, the distance threshold doff is calculated by the formula:
doff=f(d)×min(L(P0),L(Pn))
the calculation formula of the angle threshold aoff is as follows:
aoff=g(d)
wherein f, (d) and g (d) are the current processing segment P0And a segment P to be matchednFragment distance D (P) of0,Pn) As a linear function of the argument.
Preferably, in the step S58, the specific steps of initializing the parameters by the new combination element are as follows:
step A: initializing the direction vector of the combined element to the long side direction vector of the minimum circumscribed rectangle, wherein the direction is from bottom to top;
and B: initializing the front end point of the combined element to the front end point of its last sub-element;
and C: initializing a back end point of the combined element to a back end point of its first child element;
step D: and setting a crack flag of the combined element to be-1.
Preferably, in the step S6, the step of cleaning the detected crack data includes: judging whether the sum of the lengths of all the sub-elements of the combined crack in the crack data is larger than a preset threshold value l or notcIf so, retaining the combined fracture data, and if not, discarding the combined fracture data.
Compared with the prior art, the technical scheme of the invention has the beneficial effects that: by adopting a crack identification algorithm based on recursive search to identify the crack segments, the interference characteristics with equivalent size, equivalent form or equivalent color can be effectively eliminated, and the influence of the interference characteristics on crack identification is effectively eliminated; threshold setting is adopted during crack identification, so that the method is more robust; the image segmentation method adopting local dynamic threshold segmentation can effectively weaken the influence of interference factors such as illumination change, image background texture and the like on threshold segmentation.
Drawings
Fig. 1 is a flowchart of a bridge crack identification method based on recursive search according to this embodiment.
Fig. 2 is a flowchart of search identification on a fracture fragment set according to the present embodiment.
Fig. 3 is a grayscale image of a crack image to be identified according to the embodiment.
Fig. 4 is an image of the present embodiment with the local dynamic threshold segmentation completed.
FIG. 5 is a diagram illustrating a suspicious fracture fragment according to the present embodiment.
Fig. 6 is a crack image of the present embodiment after completion of the identification detection.
Fig. 7 is a simplified schematic diagram of a fracture fragment of the present embodiment.
FIG. 8 is a schematic diagram of a fracture fragment combination for completing a search according to the present embodiment.
FIG. 9 is a schematic diagram of a fracture fragment of a new element set with initialization merging completed according to the embodiment.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent;
for the purpose of better illustrating the embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product;
it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
Fig. 1 is a flowchart of a bridge crack identification method based on recursive search according to this embodiment.
The embodiment provides a bridge crack identification method based on recursive search, which comprises the following steps:
s1: and converting the crack image to be identified into a crack gray image G, and performing threshold segmentation on the crack image by adopting a local dynamic threshold segmentation method to obtain a segmentation image T.
As shown in fig. 3 and 4, a gray scale image of the crack image to be identified and a segmentation image completing local dynamic threshold segmentation in the present embodiment are respectively shown.
In this step, the specific steps of performing threshold segmentation on the crack image by using a local dynamic threshold segmentation method are as follows:
s11: carrying out mean value filtering on the crack gray level image G to obtain a mean value filtering image M;
s12: and (3) making the image T be | M-G |, traversing the image T, and when the gray value of a pixel point of the image T is greater than a preset threshold offset, setting the gray value of the pixel point to be 255 to obtain a segmented image T which is subjected to threshold segmentation.
The set threshold offset is obtained through multiple tests, and in this embodiment, the range of the set threshold offset is 2 to 8.
S2: to the aboveSegmenting the image T to carry out feature screening to obtain a suspicious crack segment set Cp={CpiAnd h, wherein i is the sequence number of the suspicious fracture fragment.
In this step, the specific steps of feature screening for the segmentation image T are as follows:
s21: detecting connected component contour C of segmented image TkWherein k represents the serial number of the detected connected domain profile;
s22: calculating the connected component contour CkThe length-width ratio f and the circularity e of the minimum circumscribed rectangle are calculated, whether suspicious crack judgment conditions are met is judged, and if yes, the connected domain contour C is subjected to classificationkMarking as suspicious fracture fragment CpiWherein the aspect ratio f and the circularity e are calculated as follows:
s23: after traversing all connected domains, marking the connected domains as suspicious crack segments CpiIntegration into a set of suspicious fracture fragments Cp={Cpi}。
Fig. 5 is a schematic diagram of a suspicious fracture fragment according to this embodiment.
In this step, the suspicious crack is judged under the condition that the connected domain contour CkAspect ratio f > t0Or connected domain profile CkThe circularity e of (a) satisfies (t)1E and f > t2Wherein, t0、t1、t2Is a predetermined threshold value obtained by a plurality of experiments, and t0>t2. In the present embodiment, t0Has a value range of (2, 4), t1Has a value range of (0.03, 0.07), t2The value range of (1) to (3).
S3: collecting C the suspicious fracture fragmentspThe parameter initialization is performed one by one.
In this step, the parameters to be initialized include fracture fragment CpiEnd point coordinates and direction ofQuantity, length, mark position, segment distance and segment included angle, wherein the suspicious crack segment set C is subjected topThe specific steps of initializing the parameters one by one are as follows:
definition of fracture fragment CpiThe front end point of (2) is the point with the minimum longitudinal coordinate value in the contour;
definition of fracture fragment CpiThe rear end point of (2) is the point with the maximum longitudinal coordinate value in the profile;
definition of fracture fragment CpiThe direction vector of (1) is the long side direction vector of the minimum circumscribed rectangle, and the direction of the direction vector is from one end where the rear endpoint is located to one end where the front endpoint is located;
definition of fracture fragment CpiLength L (C) ofpi) The length of the long side of the minimum circumscribed rectangle is the length of the long side;
definition of fracture fragment CpiThe fracture flag of (1);
definition of fracture fragment CpiAnd fracture fragment CpjFragment distance D (C) ofpi,Cpj) Is a fracture fragment CpiFront end point and fracture fragment C ofpjThe Euclidean distance between the rear end points of (a), wherein j is the sequence number of the crack segment;
definition of fracture fragment CpiAnd fracture fragment CpjAngle of segment A (C)pi,Cpj) Is a vectorAnd fracture fragment CpjAngle of direction vector of (1), wherein the vectorTo be a fracture fragment CpiStarting from the leading end of (A), and a fracture fragment CpjThe rear end points of (a) are concatenated into a vector.
S4: collecting C according to the suspicious crack segmentspThe longitudinal coordinate value of the rear end point, and the suspicious fracture segment set CpAnd re-sequencing according to the descending order to obtain the fracture fragment set P which finishes sequencing.
S5: and searching and identifying the crack segment set P through a crack identification algorithm based on recursive search, and outputting the detected crack data which completes the recursion.
Fig. 2 is a flowchart illustrating search and identification of a fracture fragment set according to this embodiment.
In this step, the specific steps of searching and identifying the fracture fragment set P by the fracture identification algorithm based on recursive search are as follows:
s51: initializing the number m of cracks as 1;
s52: taking a crack segment with the first crack flag bit flag being-1 from the crack segment set P as a currently processed segment P0Setting the crack flag bit as m, and judging the current processing segment P except the current processing segment P in the crack segment set P0If the fracture flag of other fracture fragments is equal to-1, executing the step S53; if not, jumping to execute the step S58;
s53: taking a crack segment with a second crack flag position of flag-1 from the crack segment set P as a segment P to be matchedn;
S54: calculating the currently processed segment P0And a segment P to be matchednFragment distance D (P) of0,Pn) And its length L (P)0)、L(Pn) Judging the segment distance D (P)0,Pn) Whether the distance is smaller than a preset distance threshold doff or not, if yes, executing a step S55; if not, further judging the segment distance D (P)0,Pn) Whether or not D (P) is satisfied0,Pn)<5×min(L(P0),L(Pn) If yes, jumping to the step of S56; if not, making m equal to m +1, and then jumping to execute the step S52;
s55: calculating the currently processed segment P0And a segment P to be matchednAngle of segment A (P)0,Pn) Judging the included angle A (P)0,Pn) Whether the angle is smaller than a preset angle threshold value aoff or not, if so, the segment P to be matched is determinednThe crack flag bit is set as m, and the current segment P to be matched is set asnAs the currently processed segment P0=PnTaking the next crack segment as the piece to be matchedLigand fragment Pn=Pn+1And jumping to execute the step of S57; if not, go to step S56;
s56: judging the current segment P to be matchednIf the element is the last element in the fracture fragment set P, if so, making m equal to m +1, and then jumping to execute the step S52; if not, taking the next crack segment as the segment P to be matchedn=Pn+1And executing the step of S57;
s57: judging the current segment P to be matchednIf the crack segment set P is not in the crack segment set P, jumping to execute the step S54, and if not, executing the step S58;
s58: judging whether the number m of the cracks is equal to the number of elements in the crack segment set P, if so, outputting the crack segment set P after the recurrence as recognized crack data; if not, combining the fracture fragments with the same fracture flag into a new fracture fragment, combining the corresponding fracture fragments into a new combined element in the set P, initializing the parameters of the new combined element, and then skipping to execute S51 to perform recursive search again.
The specific steps of initializing the parameters of the new combined element are as follows:
step A: initializing the direction vector of the combined element to the long side direction vector of the minimum circumscribed rectangle, wherein the direction is from bottom to top;
and B: initializing the front end point of the combined element to the front end point of its last sub-element;
and C: initializing a back end point of the combined element to a back end point of its first child element;
step D: and setting a crack flag of the combined element to be-1.
In addition, the calculation formula of the distance threshold doff in this step is:
doff=f(d)×min(L(P0),L(Pn))
the calculation formula of the angle threshold aoff is as follows:
aoff=g(d)
wherein f (d) and g (d) are as currentProcessing fragment P0And a segment P to be matchednFragment distance D (P) of0,Pn) In a specific implementation process, the distance threshold value doff and the angle threshold value aoff are dynamically adjusted as a linear function of the independent variable, so that the bridge crack identification method provided by the embodiment has higher robustness. In this embodiment, the value ranges of f (d) and g (d) are (0,5 °).
S6: cleaning the detected crack data, and judging whether the sum of the lengths of all sub-elements of the combined crack in the crack data is greater than a preset threshold value lcIf so, retaining the combined crack data and outputting the combined crack data as a recognition result, and if not, discarding the combined crack data.
In this embodiment, the threshold value lcThe value range is (200, 2000) determined by multiple experimental data.
As shown in fig. 6, the crack image of the present embodiment is the crack image of which the recognition detection is completed.
For the search and identification of the fracture fragment set by the recursive search-based fracture identification algorithm, in the specific implementation process, a simplified diagram of the fracture fragment is shown in fig. 7, where P is1~P10The fracture fragment set P is obtained according to the descending order and the sorting of the longitudinal coordinate values of the rear end points of the fracture fragments.
In the searching and identifying process, firstly, the number of cracks is initialized to m is 1, and P is taken1Setting a crack flag bit of the current processing segment to be flag m 1, and searching for the first time; get P2As the segment to be matched, since the segment P1And fragment P2Included angle of segment A (P) therebetween1,P2) Is larger than a preset angle threshold value aoff, so that the segment to be matched is selected backwards from the crack segment set P until the segment P to be matched4Satisfies the segment distance D (P)1,P4) Less than its corresponding distance threshold doff ═ f (d) x min (L (P)1),L(P4) And the included angle of the segment A (P)1,P4) Smaller than a predetermined angle threshold value aoff g (d), i.e. representing the segment P4And fragment P1Match, so segment P is4The crack flag bit is set to be flag (m) 1;
segment P4As the current processing segment, selecting a segment P backwards from the fracture segment set P5As the segment to be matched, since the segment P is1And fragment P2D (P) of the segment between4,P5) Is greater than the distance threshold value doff, therefore, the segment to be matched is selected backwards from the crack segment set P until the segment P to be matched6Satisfies the segment distance D (P)4,P6) Less than a distance threshold doff and a segment angle A (P)4,P6) Less than a predetermined angle threshold aoff, i.e. representing a segment P6And fragment P4Match and segment P4The crack flag bit is set to be flag (m) 1; in the same way, segment P6As the current processing segment, selecting a segment P backwards from the fracture segment set P7As the segment to be matched, repeating the above steps until the segment P is identified10When no matching segment is found yet, outputting a first set of combined fracture segments { P }1,P4,P6And setting m + 1-2;
starting a second search by taking a segment with the first flag bit of-1 from the fracture segment set P, wherein the segment P is1In the above step, flag is set to 1, so that segment P is taken in the search of this round2Setting the crack flag bit of the current processing segment to be flag m 2, and taking P as the current processing segment3As the fragments to be matched, repeating the searching steps until all the combined fracture fragment sets are completed, wherein the combined fracture fragment sets are respectively { P }1,P4,P6}、{P2}、{P3,P5,P8}、{P7And { P }9,P10And then combining the combined fracture fragment sets into one element, namely { P }1,P4,P6Combine into new element { { P { (P) }1_1,P1_2,P1_3Will { P } }, will { P2As a new element { P }2Will { P }3,P5,P8Combine into new element { { P { (P) }3_1,P3_2,P3_3Will { P } }, will { P7As a new element { P }4Will { P }9,P10Combine into new element { { P { (P) }5_1,P5_2As shown in fig. 9, it is a schematic diagram of combining new elements of the present embodiment to complete merging, and then { { P } is combined into new elements1_1,P1_2,P1_3}}、{P2}、{{P3_1,P3_2,P3_3}}、{P4}、{{P5_1,P5_2H } are merged into a new set of fracture fragments Pnew={{P1,P4,P6},{P2},{P3,P5,P8},{P7},{P9,P10And initializing a direction vector, a front end point and a rear end point of each combination element, as shown in fig. 8 and 9, which are a fracture fragment combination diagram and an initialized combination element parameter diagram for completing one search in this embodiment, respectively. If the fracture fragment is set PnewIf the number of elements in the set P is less than the number of elements in the set P, let P be PnewAnd then repeating the steps on the set P to complete the recursive search of all combined elements, otherwise ending the search and identification on the fracture fragment set P, cleaning the detected and identified fracture and outputting the fracture.
In the embodiment, the crack segments are identified by adopting a crack identification algorithm based on recursive search, so that the interference characteristics with equivalent size, equivalent form or equivalent color can be effectively eliminated, and the influence of the interference characteristics on crack identification is effectively eliminated; threshold setting is adopted during crack identification, so that the method is more robust; the image segmentation method adopting local dynamic threshold segmentation can effectively weaken the influence of interference factors such as illumination change, image background texture and the like on threshold segmentation.
The same or similar reference numerals correspond to the same or similar parts;
the terms describing positional relationships in the drawings are for illustrative purposes only and are not to be construed as limiting the patent;
it should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (5)
1. A bridge crack identification method based on recursive search is characterized by comprising the following steps:
s1: converting a crack image to be identified into a crack gray image G, and performing threshold segmentation on the crack image by adopting a local dynamic threshold segmentation method to obtain a segmentation image T;
s2: performing characteristic screening on the segmentation image T to obtain a suspicious crack segment set Cp={CpiH, wherein i is the sequence number of the suspicious crack segment;
the specific steps of performing feature screening on the segmented image T are as follows:
s21: detecting connected component contour C of segmented image TkWherein k represents the serial number of the detected connected domain profile;
s22: calculating the connected component contour CkThe length-width ratio f and the circularity e of the minimum circumscribed rectangle are calculated, whether suspicious crack judgment conditions are met is judged, and if yes, the connected domain contour C is subjected to classificationkMarking as suspicious fracture fragment CpiWherein the aspect ratio f and the circularity e are calculated as follows:
wherein the suspicious crack judgment condition is that the connected domain profile CkAspect ratio f > t0Or connected domain profile CkThe circularity e of (a) satisfies (t)1E and f > t2Wherein, t0、t1、t2Is a predetermined threshold value obtained by a plurality of experiments, and t0>t2;
S23: after traversing all connected domains, marking the connected domains as suspicious crack segments CpiIntegration into a set of suspicious fracture fragments Cp={Cpi};
S3: collecting C the suspicious fracture fragmentspInitializing parameters one by one, wherein the parameters comprise endpoint coordinates of suspicious fracture fragments; wherein: the initialization parameters also include fracture fragment CpiThe direction vector, the length, the mark position, the segment distance and the segment included angle of the fracture, wherein the suspicious fracture segment set C is collectedpThe specific steps of initializing the parameters one by one are as follows:
definition of fracture fragment CpiThe front end point of (2) is the point with the minimum longitudinal coordinate value in the contour;
definition of fracture fragment CpiThe rear end point of (2) is the point with the maximum longitudinal coordinate value in the profile;
definition of fracture fragment CpiThe direction vector of (1) is the long side direction vector of the minimum circumscribed rectangle, and the direction of the direction vector is from one end where the rear endpoint is located to one end where the front endpoint is located;
definition of fracture fragment CpiLength L (C) ofpi) The length of the long side of the minimum circumscribed rectangle is the length of the long side;
definition of fracture fragment CpiThe fracture flag of (1);
definition of fracture fragment CpiAnd fracture fragment CpjFragment distance D (C) ofpi,Cpj) Is a fracture fragment CpiFront end point and fracture fragment C ofpjThe Euclidean distance between the rear end points of (a), wherein j is the sequence number of the crack segment;
definition of fracture fragment CpiAnd fracture fragment CpjAngle of segment A (C)pi,Cpj) Is a vectorAnd fracture fragment CpjAngle of direction vector of (1), wherein the vectorTo be crackedFragment CpiStarting from the leading end of (A), and a fracture fragment CpjThe rear end points of (1) are connected into a vector;
s4: collecting C according to the suspicious crack segmentspThe longitudinal coordinate value of the rear end point, and the suspicious fracture segment set CpReordering according to the descending order to obtain a crack segment set P which finishes sequencing;
s5: outputting the identified crack data which completes the recursion to a crack segment set P through a crack identification algorithm based on the recursion search; the specific steps of searching and identifying the fracture fragment set P through the fracture identification algorithm based on recursive search are as follows:
s51: initializing the number m of cracks as 1;
s52: taking a crack segment with the first crack flag bit flag being-1 from the crack segment set P as a currently processed segment P0Setting the crack flag bit as m, and judging the current processing segment P except the current processing segment P in the crack segment set P0If the fracture flag of other fracture fragments is equal to-1, executing the step S53; if not, jumping to execute the step S58;
s53: taking a crack segment with a second crack flag position of flag-1 from the crack segment set P as a segment P to be matchedn;
S54: calculating the currently processed segment P0And a segment P to be matchednFragment distance D (P) of0,Pn) And its length L (P)0)、L(Pn) Judging the segment distance D (P)0,Pn) Whether the distance is smaller than a preset distance threshold doff or not, if yes, executing a step S55; if not, further judging the segment distance D (P)0,Pn) Whether or not D (P) is satisfied0,Pn)<5×min(L(P0),L(Pn) If yes, jumping to the step of S56; if not, making m equal to m +1, and then jumping to execute the step S52;
s55: calculating the currently processed segment P0And a segment P to be matchednAngle of segment A (P)0,Pn) Judging the included angle A (P)0,Pn) Whether or not it is less than a preset angle threshold aoffIf yes, the segment P to be matched is determinednThe crack flag bit is set as m, and the current segment P to be matched is set asnAs the currently processed segment P0=PnTaking the next crack segment as the segment P to be matchedn=Pn+1And jumping to execute the step of S57; if not, go to step S56;
s56: judging the current segment P to be matchednIf the element is the last element in the fracture fragment set P, if so, making m equal to m +1, and then jumping to execute the step S52; if not, taking the next crack segment as the segment P to be matchedn=Pn+1And executing the step of S57;
s57: judging the current segment P to be matchednIf the crack segment set P is not in the crack segment set P, jumping to execute the step S54, and if not, executing the step S58;
s58: judging whether the number m of the cracks is equal to the number of elements in the crack segment set P, if so, outputting the crack segment set P after the recurrence as recognized crack data; if not, combining the crack segments with the same crack flag into a new crack segment, combining the corresponding crack segments into a new combined element in the set P, initializing the parameters of the new combined element, and then skipping to execute S51 to perform recursive search again;
s6: and cleaning the detected crack data and outputting the data as a recognition result.
2. The bridge crack identification method of claim 1, wherein: in the step S1, the specific steps of performing threshold segmentation on the crack image by using a local dynamic threshold segmentation method are as follows:
s11: carrying out mean value filtering on the crack gray level image G to obtain a mean value filtering image M;
s12: and (4) setting the image T as M-G, traversing the image T, and when the gray value of a pixel point of the image T is greater than a preset threshold offset, setting the gray value of the pixel point to be 255 to obtain a segmented image T which is subjected to threshold segmentation.
3. The bridge crack identification method of claim 1, wherein: the calculation formula of the distance threshold doff is as follows:
doff=f(d)×min(L(P0),L(Pn))
the calculation formula of the angle threshold aoff is as follows:
aoff=g(d)
wherein f, (d) and g (d) are the current processing segment P0And a segment P to be matchednFragment distance D (P) of0,Pn) As a linear function of the argument.
4. The bridge crack identification method of claim 1, wherein: in the step S58, the specific steps of initializing the parameters of the new combination element are as follows:
step A: initializing the direction vector of the combined element to the long side direction vector of the minimum circumscribed rectangle, wherein the direction is from bottom to top;
and B: initializing the front end point of the combined element to the front end point of its last sub-element;
and C: initializing a back end point of the combined element to a back end point of its first child element;
step D: and setting a crack flag of the combined element to be-1.
5. The bridge crack identification method of claim 1, wherein: in the step S6, the specific step of cleaning the detected crack data includes: judging whether the sum of the lengths of all the sub-elements of the combined crack in the crack data is larger than a preset threshold value l or notcIf so, retaining the combined fracture data, and if not, discarding the combined fracture data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910673031.6A CN110487497B (en) | 2019-07-24 | 2019-07-24 | Bridge crack identification method based on recursive search |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910673031.6A CN110487497B (en) | 2019-07-24 | 2019-07-24 | Bridge crack identification method based on recursive search |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110487497A CN110487497A (en) | 2019-11-22 |
CN110487497B true CN110487497B (en) | 2021-09-24 |
Family
ID=68548207
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910673031.6A Active CN110487497B (en) | 2019-07-24 | 2019-07-24 | Bridge crack identification method based on recursive search |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110487497B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111105408B (en) * | 2019-12-27 | 2022-05-03 | 南京大学 | Building surface crack detection method and system based on image processing |
CN111784667B (en) * | 2020-06-30 | 2023-09-26 | 京东科技信息技术有限公司 | Crack identification method and device |
CN116630899B (en) * | 2023-07-21 | 2023-10-20 | 四川公路工程咨询监理有限公司 | Highway side slope disease monitoring and early warning system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104008553B (en) * | 2014-06-17 | 2017-03-22 | 武汉武大卓越科技有限责任公司 | Crack detection method with image gradient information and watershed method conflated |
-
2019
- 2019-07-24 CN CN201910673031.6A patent/CN110487497B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN110487497A (en) | 2019-11-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110487497B (en) | Bridge crack identification method based on recursive search | |
US11922615B2 (en) | Information processing device, information processing method, and storage medium | |
CN110678901B (en) | Information processing apparatus, information processing method, and computer-readable storage medium | |
CN108280823B (en) | Method and system for detecting weak edge flaws on optical cable surface in industrial production | |
CN106372642B (en) | Ellipse rapid detection method based on contour curve segmentation arc merging and combining | |
CN102426649B (en) | Simple steel seal digital automatic identification method with high accuracy rate | |
CN116721106A (en) | Profile flaw visual detection method based on image processing | |
CN109711418B (en) | Contour corner detection method for object plane image | |
CN114418957A (en) | Global and local binary pattern image crack segmentation method based on robot vision | |
CN115082419A (en) | Blow-molded luggage production defect detection method | |
CN115294159B (en) | Method for dividing corroded area of metal fastener | |
CN115082466B (en) | PCB surface welding spot defect detection method and system | |
JP2017072517A (en) | Detection method of surface crack of concrete | |
CN113989266B (en) | Mechanical part surface defect detection method based on image processing | |
CN110427979B (en) | Road water pit identification method based on K-Means clustering algorithm | |
CN111080631A (en) | Fault positioning method and system for detecting floor defects of spliced images | |
CN110749598A (en) | Silkworm cocoon surface defect detection method integrating color, shape and texture characteristics | |
CN112991283A (en) | Flexible IC substrate line width detection method based on super-pixels, medium and equipment | |
KR102195940B1 (en) | System and Method for Detecting Deep Learning based Human Object using Adaptive Thresholding Method of Non Maximum Suppression | |
CN116486091B (en) | Fan blade defect area rapid segmentation method and system based on artificial intelligence | |
CN106897723B (en) | Target real-time identification method based on characteristic matching | |
CN111340134B (en) | Rapid template matching method based on local dynamic warping | |
JP2000339462A (en) | Defect inspection method and defect inspection device | |
CN111105408B (en) | Building surface crack detection method and system based on image processing | |
CN110060270B (en) | Edge detection method suitable for polygonal tubular object with low imaging quality |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |