CN107179322A - A kind of bridge bottom crack detection method based on binocular vision - Google Patents
A kind of bridge bottom crack detection method based on binocular vision Download PDFInfo
- Publication number
- CN107179322A CN107179322A CN201710452913.0A CN201710452913A CN107179322A CN 107179322 A CN107179322 A CN 107179322A CN 201710452913 A CN201710452913 A CN 201710452913A CN 107179322 A CN107179322 A CN 107179322A
- Authority
- CN
- China
- Prior art keywords
- mtd
- mrow
- msub
- mtr
- mtable
- 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
Classifications
-
- 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
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Processing (AREA)
Abstract
The invention discloses kind of a bridge bottom crack detection method, the full-size(d) in crack is reduced using binocular vision, error is greatly evaded, avoid in monocular vision, because the video camera camera plane in monocular vision and bridge bottom surface are not parallel, it is only projection of the crack in monocular-camera camera plane to cause captured obtained crack pattern picture, in monocular vision, simple image procossing can be passed through to such crack picture, what is so calculated misses by a mile, the size of projection of the crack in monocular-camera camera plane is only obtained, it is not crack full-size(d), in terms of image procossing, present invention employs improved medium filtering, compared to traditional medium filtering, intermediate value replacement will be carried out to pixel all in image by avoiding, the noise detected is only carried out intermediate value replacement, the method more remains the detailed information in crack in image, avoid afterwards making after filtering crack pattern as excess smoothness.
Description
Technical field
The present invention relates to bridge machinery field, more particularly to a kind of bridge bottom crack detection side based on binocular vision
Method.
Background technology
In recent years, with the fast development of China, the science of bridge building of China has also obtained huge development, the bridge of China
Total kilometrage has reached a thrilling mileage, and at the same time with the increase of car ownership, bridge is used as the one of traffic
Part, has also played huge traffic pressure with regard to unavoidable commitment.There is significant component of overloaded vehicle in China, this is right
The technology status of bridge proposes a very big requirement, occurs in that many bridges do not reach service life, just becomes unsafe bridge,
And quantity remains high always, the technology status of bridge directly threatens the life security of people, and this causes the technology of bridge
Situation becomes focus of concern.
Only by increasing the dynamics and frequency that are detected to Bridge Crack, it can ensure that the technology status of bridge is stable,
When carrying out the durability evaluating of bridge, the dimensional parameters of Bridge Crack are an important reference aspects, to bridge inspection and maintenance portion
For door, this can provide reliable foundation, so that maintenance department timely can repair to bridge, it is ensured that it is pacified
Quan Xing.But in China, the detection of current bridge bottom crack is also to rely on manpower progress, relies primarily on big machinery or bridge
Bridge machinery expert is sent to bridge bottom surface by beam detection car, is entered by bridge machinery expert using the flaw size measuring instrument such as scale
The lookup of pedestrian's work simultaneously measures crack.This just inevitably brings the defect of the following aspects.It is to measure into first
Present aspect, spends too high.The expense for renting big machinery or bridge inspection vehicle be one day thousands of even tens of thousands of pieces, and also need
Expert is asked, it is higher that this allows for manpower expense.Second be precision aspect, is fixed against Bridge Crack detection expert one by one
Crack is found, this obviously can cause the intervention of subjective factor, so as to have influence on the accuracy of crack searching, manually carried out with instrument
The measurement of flaw size, this random error that certainly will produce reading causes the problem of precision is low.In terms of 3rd is detection efficiency, allow
Bridge Crack detects that expert goes to searching crack one by one to bridge bottom, and this is clearly the thing of a time-consuming effort again so that
Detection time to a bridge generally requires to complete some months.4th is the safety problem of expert, and expert measures at bridge bottom
During crack, it may occur that some are dangerous.
Recent years, scholars both domestic and external in the picture for attempting to utilize monocular camera to shoot a large amount of bridge bottom surfaces,
Then fracture picture carries out gray processing, and the simple image procossing such as smoothing denoising obtains the pixel count in crack, afterwards using splitting
The pixel count occupied in picture and camera pixel rate (unit pixel size) product are sewn on, so as to draw the size in crack.This side
Though method is than manually operating simpler, safety, monocular camera is in shooting process it is difficult to ensure that the camera plane of video camera
It is parallel with bridge bottom surface, thus captured obtained picture, often projection of the bridge bottom surface in video camera camera plane, from
And resulting crack pattern picture is not its real size, it is impossible to reflect its full-size(d).So by means of which
The flaw size error arrived is big, and precision is than relatively low.
The content of the invention
It is existing to overcome it is an object of the invention to provide a kind of bridge bottom crack detection method based on binocular vision
The deficiency of technology.
To reach above-mentioned purpose, the present invention is adopted the following technical scheme that：
A kind of bridge bottom crack detection method based on binocular vision, specifically includes following steps：
1) binocular vision dual image collection, is carried out to bridge bottom first；
2), using weighted mean method by step 1) obtained dual image crack image gray processing, then pass through medium filtering
Denoising is carried out, image enhaucament has been carried out using the piecewise linear function of selected threshold value, edge of crack has been carried out using Sobel operators
Extract, finally obtain the binary map of crack pattern picture；
3), to step 1) dual image of collection demarcated by Zhang Zhengyou standardizations, then to step 2) obtained two-value
Figure carries out the matching of the binary map of gained crack pattern picture using the indeformable description operators of Fourier-Mellin, finally by by picture
Point under plain coordinate system is converted to the point under world coordinate system, and the length and width in crack have been calculated using Euclidean distance formula
Degree.
Further, step 1) in, dual image collection is carried out using UAV flight's binocular camera, binocular camera is put down
Row is arranged on unmanned plane, assist illuminator is equipped with above unmanned plane, for illuminating bridge bottom surface.
Further, step 2) in, dual image crack image gray processing：By obtained dual image crack imagery exploitation weighting
The method of average, is to be weighted average computation using following formula and obtain ash the component of the red R of coloured image, green G, indigo plant B triple channels
Degree figure；
F (x, y)=0.299R (x, y)+0.587G (x, y)+0.114B (x, y) (1)
In above formula, f (x, y) is the gray value of image slices vegetarian refreshments, and R (x, y), G (x, y), B (x, y) are respectively coloured image
Red R, green G, the component value of indigo plant B triple channels of pixel.
Further, medium filtering denoising：Pending pixel, is placed on window by the filter window that make use of size to be 3*3
Mouth center, obtains the maximum and minimum value of the gray value for all pixels point that filter window is included, if judging center pixel
The gray value of point is equal to gray scale maximum or minimum value, then it is assumed that this pixel is noise spot, then using in the gray scale in window
Value is replaced, if the gray value of central pixel point is not equal to gray scale maximum or minimum value, then it is assumed that this pixel is signal
Point, the gray value of itself is used as output.
Further, image enhaucament：The crack picture of above-mentioned medium filtering denoising is utilized as given by following formula
The piecewise linear function of selected threshold value carries out the brightness of specific tonal range in image enhancement processing, prominent image, distinguishes significantly
Crack and the color of background；
In above formula, G represents the gray value after pixel conversion, gmax、gminThe gray scale of pixel in entire image is represented respectively
Maximum and minimum value, g represent the gray value before pixel conversion；With reference to enhancing effect, selected threshold is gmax=0.3, gmin=
0.7。
Further, Sobel operators carry out edge extracting：Above-mentioned crack picture is subjected to edge inspection using Sobel operators
Survey, the gray value f (x, y) of each pixel in image is subjected to convolution algorithm with the two convolution kernels respectively, maximum is taken
Value is as output, and what is obtained after computing is the image that a width embodies edge amplitude.
Further, step 3) in, binocular camera is demarcated by Zhang Zhengyou standardizations, using Fourier-
The indeformable description operators of Mellin carry out the matching of crack binary map obtained by binocular camera, finally by by under pixel coordinate system
Point be converted to point under world coordinate system, the length and width in crack has been calculated using Euclidean distance formula.
Further, binocular camera is demarcated：Three dimensional space coordinate is as follows relative to the corresponding relation of two-dimensional coordinate：
In above formula, M is the internal reference matrix of binocular camera, wherein lx、lyRespectively binocular camera is in horizontal and vertical side
Unit pixel size on (x and y directions), ey、eyRespectively binocular camera is on both horizontally and vertically (x and y directions)
Distortion factor, a is scale factor, and V is mapping matrix of the binocular camera coordinate system relative to world coordinate system, and wherein U is 3
Rank spin matrix, c is 3*1 translation vector, and equation high order end is two-dimensional coordinate, and low order end is three dimensional space coordinate；Zhang Zhengyou
What standardization was carried out is plane reference, therefore makes scaling board be in Z=0 plane, then above formula is changed into：
Wherein, the mapping matrix of three dimensional space coordinate to two-dimensional coordinate is H；
Position of the darkened features point under pixel coordinate system is recognized and calculated by image procossing and obtained in scaling board, and its
World coordinates can be obtained by scaling board, as available from the above equation：
And then obtain：
Spin matrix U is unit orthogonal matrix, therefore column vector is orthogonal and is unit vector, then has：
Therefore all there are following restriction relations to the internal reference of calibration for cameras for any piece image, simultaneous formula (6),
(7) it can obtain：
Projection matrix H is drawn by above formula.
Further, images match：Two width binary maps of above-mentioned treated binocular camera are subjected to images match, it is false
If two images f1(x, y) and f2(x, y) has rotation, scaling, the relation of translation, then the relation between them can be represented
It is as follows：
f2(x, y)=f1(a(xcosβ+ysinβ)-Δx,a(-xsinβ+ycosβ)-Δy) (9)
Wherein, Δx、ΔyThe respectively horizontally and vertically translation vector on (x and y), a is the ratio between two width figures
Example zoom factor, is the anglec of rotation between two width figures；Carrying out Fourier transformation can obtain：
Wherein, f2The spectrum phase of (x, y) isIt is relevant with the anglec of rotation, translational movement and zoom factor, right
(10) Shi Mo get power spectrum：
|F2(m, n) |=a-2|F1(a-1(mcosβ+nsinβ),a-1(-msinβ+ncosβ))| (11)
From (10) formula and (11) formula, image has zoom factor a, then its power spectrum just has zoom factor a-1, will scheme
As rotation, its anglec of rotation is β, then its power spectrum will rotate identical angle, and for spectrum center (m=n=0), it is to rotation
Gyration and yardstick are all constant, due to Δx、ΔyZoom factor and the anglec of rotation are determined, then is transformed to a, β translate shape
Formula, is first converted to polar coordinates by frequency spectrum：
Make hypothesis below：
Then by (12) Shi Ke get：
Sρ(β1, ρ) and=a-2Rρ((β1-β),ρ/a) (14)
λ=log ρ, k=loga are set simultaneously, then (14) formula can be deformed into：
Sρl(β1, ρ) and=a-2Rρl((β1-β),λ-k) (15)
(15) in formula, SρlRepresent logarithmic transformation, RρlIt is the indeformable description operators of Fourier-Mellin；Become after conversion
For：
Sρl(m, n)=a-2Rρl(m,n)exp(-2jπ(mβ+nk)) (16)
Thus (16) formula can be converted to the gap of the anglec of rotation of two images and zoom factor the gap of translational movement, right
In the gap of translational movement, tried to achieve using the Fourier inversion of crosspower spectrum.
Further, the calculating of fracture width and length：Obtained by images match after one group of pixel matched, i.e.,
Binocular camera shoots the pixel in obtained crack picture, and the projection matrix drawn is demarcated using binocular camera, just may be used
Draw the pixel matched corresponding true point in three dimensions, two coordinates truly put imputed out be respectively (X,
Y, Z), (P, Q, R),
The length and width in crack is then just can obtain using above formula distance between two points formula.
Compared with prior art, the present invention has following beneficial technique effect：
A kind of bridge bottom crack detection method based on binocular vision of the present invention, crack is reduced using binocular vision
Full-size(d), has greatly evaded error, it is to avoid in monocular vision, due to the video camera camera plane and bridge in monocular vision
Soffit is not parallel, and it is only projection of the crack in monocular-camera camera plane to cause captured obtained crack pattern picture,
, can be to such crack picture by simple image procossing in monocular vision, the picture occupied in the picture using crack
The product of prime number and camera pixel rate calculates the size for obtaining crack, and what is so calculated misses by a mile, and is only split
The size for the projection being sewn in monocular-camera camera plane, is not crack full-size(d), and in terms of image procossing, the present invention is adopted
With improved medium filtering, compared to traditional medium filtering, it is to avoid during pixel all in image will be carried out
Value is replaced, and the noise detected is only carried out intermediate value replacement, and the method more remains the detailed information in crack in image, kept away
Exempt from afterwards to make after filtering crack pattern as excess smoothness.
Further, using UAV flight's binocular camera, so that the problem in artificial detection crack is solved, and profit
The full-size(d) in crack can be reduced with binocular vision, it is possible to use the point under pixel coordinate system is calculated under its world coordinate system
True point coordinate so as to solve monocular-camera shoot precision it is low the problem of, in terms of image procossing, the present invention use
Improved medium filtering, compared to traditional medium filtering, the method more remains the detailed information in crack in image,
Avoid making crack pattern as excess smoothness afterwards after filtering, the piecewise linearity for choosing appropriate threshold is make use of in terms of image enhaucament
Function greatly improves the contrast of crack and background.In addition using four rotor wing unmanned aerial vehicles greatly improve manpower shoot and
The limitation of traditional monocular-camera shooting angle, it is motor-driven due to unmanned plane itself using UAV flight's binocular camera
Property, it is possible to realize shooting to bridge bottom more perspective, it is to avoid when conventional method is shot, the shooting angle in some regions
Difficult the problem of.
Further, using the projection matrix obtained by being demarcated to binocular camera, pixel m and n can just be calculated alive
The true point M in corresponding crack coordinate under boundary's coordinate system, the coordinate that can be truly put in crack using the method is calculated, and is utilized
Distance between two points formula just can draw the full-size(d) in crack, so as to evade the significant errors that monocular vision occurs.
Brief description of the drawings
Fig. 1 is the flow chart of the inventive method.
Fig. 2 is the flow chart that binocular vision of the present invention is calculated.
Fig. 3 is gridiron pattern scaling board.
Fig. 4 is the graph of a relation between image coordinate system, pixel coordinate system, camera coordinate system and world coordinate system.
Fig. 5 is piecewise linear function.
Fig. 6 is binocular vision imaging figure.
Fig. 7 is monocular vision image.
Embodiment
The present invention is described in further detail below in conjunction with the accompanying drawings：
As shown in figure 1, a kind of bridge bottom crack detection method based on binocular vision, specifically includes following steps：
4) binocular vision dual image collection, is carried out to bridge bottom first；
5), using weighted mean method by step 1) obtained dual image crack image gray processing, then pass through medium filtering
Denoising is carried out, image enhaucament has been carried out using the piecewise linear function of selected threshold value, edge of crack has been carried out using Sobel operators
Extract, finally obtain the binary map of crack pattern picture；
6), to step 1) dual image of collection demarcated by Zhang Zhengyou standardizations, then to step 2) obtained two-value
Figure carries out the matching of the binary map of gained crack pattern picture using the indeformable description operators of Fourier-Mellin, finally by by picture
Point under plain coordinate system is converted to the point under world coordinate system, and the length and width in crack have been calculated using Euclidean distance formula
Degree.
Step 1) in, dual image collection is carried out using UAV flight's binocular camera, binocular camera is set in parallel in
On unmanned plane, assist illuminator is equipped with above unmanned plane, for illuminating bridge bottom surface.
Step 2) in, dual image crack image gray processing：By obtained dual image crack imagery exploitation weighted mean method, i.e.,
It is that the component of the red R of coloured image, green G, indigo plant B triple channels is weighted average computation using following formula and obtains gray-scale map；
F (x, y)=0.299R (x, y)+0.587G (x, y)+0.114B (x, y) (1)
In above formula, f (x, y) is the gray value of image slices vegetarian refreshments, and R (x, y), G (x, y), B (x, y) are respectively coloured image
Red R, green G, the component value of indigo plant B triple channels of pixel；
Medium filtering denoising：Pending pixel, is placed on window center, obtained by the filter window that make use of size to be 3*3
The maximum and minimum value of the gray value of all pixels point included to filter window, if judging the gray value of central pixel point
Equal to gray scale maximum or minimum value, then it is assumed that this pixel is noise spot, then is replaced using the gray scale intermediate value in window,
If the gray value of central pixel point is not equal to gray scale maximum or minimum value, then it is assumed that this pixel is signaling point, the ash of itself
Angle value is used as output；
Image enhaucament：The crack picture of above-mentioned medium filtering denoising is utilized as selected threshold value given by following formula
Piecewise linear function carries out the brightness of specific tonal range in image enhancement processing, prominent image, and crack and background are distinguished significantly
Color；
In above formula, G represents the gray value after pixel conversion, gmax、gminThe gray scale for representing pixel in entire image is maximum
Value and minimum value, g represent the gray value before pixel conversion；With reference to enhancing effect, selected threshold is gmax=0.3, gmin=0.7.
The figure of piecewise linear function as shown in Figure 5.
Sobel operators carry out edge extracting：In order to avoid excessively refining, many disturbing factors are caused also to be treated as crack side
Edge and be extracted, by above-mentioned crack picture utilize Sobel operators carry out rim detection, Sobel horizontally and vertically
Operator template is as shown in table 1, table 2, and table 1 is that horizontal direction template is used for the edge of crack in detection level direction, and table 2 is Vertical Square
To template be used to detect the edge of crack of vertical direction, the gray value f (x, y) of each pixel in image is used respectively
The two convolution kernels carry out convolution algorithm, take maximum as output, what is obtained after computing is that a width embodies edge width
The image of degree.
Table 1
| -1 | -2 | -1 |
| 0 | 0 | 0 |
| 1 | 2 | 1 |
Table 2
| -1 | 0 | 1 |
| -2 | 0 | 2 |
| -1 | 0 | 1 |
Step 3) in, binocular camera is demarcated by Zhang Zhengyou standardizations, it is constant using Fourier-Mellin
Shape describes the matching that operator carries out crack binary map obtained by binocular camera, is converted to finally by by the point under pixel coordinate system
Point under world coordinate system, the length and width in crack has been calculated using Euclidean distance formula.
As shown in Fig. 2 flow charts, binocular camera demarcation：In three dimensions, analysis position is to be unable to do without to take the photograph with scene
Camera coordinate system, image coordinate system, pixel coordinate system and world coordinate system, world coordinate system is：3 d space coordinate system, 4
The relation of individual coordinate system is as shown in Figure 4.
What the demarcation of binocular camera referred to is just to determine that three-dimensional coordinate, to the projection matrix of two-dimensional coordinate, has obtained projection square
After battle array, you can obtain the corresponding relation of pixel coordinate and three-dimensional coordinate, binocular camera after demarcation obtained by it is inside and outside
Parameter determines projection matrix, and outer ginseng refers to that camera coordinate system is relative to the translation coefficient of three-dimensional system of coordinate and rotation
Number, internal reference refers to the distortion factor of binocular camera, and Zhang Zhengyou standardizations scaling board is gridiron pattern scaling board, gridiron pattern demarcation
Plate is as shown in figure 3, three dimensional space coordinate is as follows relative to the corresponding relation of two-dimensional coordinate：
In above formula, M is the internal reference matrix of binocular camera, wherein lx、lyIt is binocular camera in both horizontally and vertically (x
With y directions) on unit pixel size, ey、eyFor distortion system of the binocular camera on both horizontally and vertically (x and y directions)
Number, a is scale factor, and V is mapping matrix of the binocular camera coordinate system relative to world coordinate system, and wherein U is 3 rank spin moments
Battle array, c is 3*1 translation vector, and equation high order end is two-dimensional coordinate, and low order end is three dimensional space coordinate；Zhang Zhengyou standardizations enter
Capable is plane reference, therefore makes scaling board be in Z=0 plane, then above formula is changed into：
Wherein, the mapping matrix of three dimensional space coordinate to two-dimensional coordinate is H；
Position of the darkened features point under pixel coordinate system can be recognized and be calculated by image procossing and obtained in scaling board,
And its world coordinates can be obtained by scaling board；Accordingly, obtained parameter is needed to be five using Zhang Zhengyou standardizations.By above formula
It can obtain：
And then obtain：
Spin matrix U is unit orthogonal matrix, therefore column vector is orthogonal and is unit vector, then has：
Therefore all there are following restriction relations to the internal reference of calibration for cameras for any piece image, simultaneous formula (6),
(7) it can obtain：
It follows that projection matrix H can just be drawn by shooting minimum three pictures, and the interior of binocular camera can be drawn
Outer parameter.
Images match：Two width binary maps of above-mentioned treated left and right cameras are subjected to images match, are employed herein
The indeformable description operators of Fourier-Mellin, this is the improvement for Fourier transformation, greatly improves efficiency, is saved
Time, and the precision of matching has been obtained strong guarantee.Assuming that two images f1(x, y) and f2(x, y) have rotation,
Scaling, the relation of translation, then the relation between them can be expressed as follows：
f2(x, y)=f1(a(xcosβ+ysinβ)-Δx,a(-xsinβ+ycosβ)-Δy) (9)
Wherein, Δx、ΔyFor the translation vector on horizontally and vertically (x and y), a is the ratio contracting between two width figures
The factor is put, is the anglec of rotation between two width figures；Carrying out Fourier transformation can obtain：
Wherein, f2The spectrum phase of (x, y) isIt is relevant with the anglec of rotation, translational movement and zoom factor, right
(10) Shi Mo get power spectrum：
|F2(m, n) |=a-2|F1(a-1(mcosβ+nsinβ),a-1(-msinβ+ncosβ))| (11)
From (10) formula and (11) formula, image has zoom factor a, then its power spectrum just has zoom factor a-1, will scheme
As rotation, its anglec of rotation is β, then its power spectrum will rotate identical angle, and for spectrum center (m=n=0), it is to rotation
Gyration and yardstick are all constant, due to Δx、ΔyZoom factor and the anglec of rotation are determined, then is transformed to a, β translate shape
Formula, is first converted to polar coordinates by frequency spectrum：
Make hypothesis below：
Then by (12) Shi Ke get：
Sρ(β1, ρ) and=a-2Rρ((β1-β),ρ/a) (14)
λ=log ρ, k=loga are set simultaneously, then (14) formula can be deformed into：
Sρl(β1, ρ) and=a-2Rρl((β1-β),λ-k) (15)
(15) in formula, SρlRepresent logarithmic transformation, RρlIt is the indeformable description operators of Fourier-Mellin；Become after conversion
For：
Sρl(m, n)=a-2Rρl(m,n)exp(-2jπ(mβ+nk)) (16)
Thus (16) formula can be converted to the gap of the anglec of rotation of two images and zoom factor the gap of translational movement, right
In the gap of translational movement, it is possible to use the Fourier inversion of crosspower spectrum is tried to achieve.
The calculating of fracture width and length：Obtained by images match after one group of pixel matched, i.e. binocular camera shooting
Machine shoots the pixel in obtained crack picture, and the projection matrix drawn is demarcated using binocular camera, phase just can be drawn
The pixel matched somebody with somebody corresponding true point in three dimensions, two coordinates truly put imputed out are respectively (X, Y, Z), (P,
Q, R), then the length and width in crack, such as formula (17) are just can obtain using distance between two points formula.
Monocular vision image as shown in Figure 7, bridge plane ABCD and monocular-camera camera plane EFCD it is not parallel and
Angle is φ, then crack HG (assuming that crack is straight line) being projected as in monocular-camera camera plane EFCD on plane ABCD
H1G, if H points and H1Point is overlapped on two plane intersection line CD, and φ=∠ GHG1, then H1G=HGcos φ, therefore deduce that and work as
It is that angle is 0 ° of ＜ φ≤90 ° during φ ≠ 0 °, H can be obtained1G ＜ HG, i.e., when plane ABCD is not parallel to plane EFCD, crack
Full-size(d) is not the flaw size on image obtained by monocular-camera is shot, and φ is bigger, HG1Smaller, error is bigger.
Binocular vision imaging figure as shown in Figure 6, it is assumed that the pixel coordinate system o of left side video camera1-x1y1z1In crack point m
With the pixel coordinate system o of the right video camera2-x2y2z2Under crack point n be a pair of images match points, using to binocular camera
Projection matrix (corresponding relation of the three dimensional space coordinate (world coordinates) relative to two-dimensional coordinate (pixel coordinate)) obtained by demarcation,
Just pixel m and n true point M in corresponding crack under world coordinate system coordinate can be calculated, will can be split using the method
Stitch true point coordinates to calculate, the full-size(d) in crack just can be drawn using distance between two points formula, so as to evade monocular
The significant errors that vision occurs.
Claims (10)
1. a kind of bridge bottom crack detection method based on binocular vision, it is characterised in that specifically include following steps：
1) binocular vision dual image collection, is carried out to bridge bottom first；
2), using weighted mean method by step 1) obtained dual image crack image gray processing, then carried out by medium filtering
Denoising, image enhaucament has been carried out using the piecewise linear function of selected threshold value, and having carried out edge of crack using Sobel operators carries
Take, finally obtain the binary map of crack pattern picture；
3), to step 1) dual image of collection demarcated by Zhang Zhengyou standardizations, then to step 2) obtained binary map adopts
The matching of the binary map of gained crack pattern picture is carried out with the indeformable description operators of Fourier-Mellin, is sat finally by by pixel
Point under mark system is converted to the point under world coordinate system, and the length and width in crack has been calculated using Euclidean distance formula.
2. a kind of bridge bottom crack detection method based on binocular vision according to claim 1, it is characterised in that step
It is rapid 1) in, dual image collection is carried out using UAV flight's binocular camera, binocular camera is set in parallel on unmanned plane, nothing
Man-machine top is equipped with assist illuminator, for illuminating bridge bottom surface.
3. a kind of bridge bottom crack detection method based on binocular vision according to claim 1, it is characterised in that step
It is rapid 2) in, dual image crack image gray processing：It is by cromogram by obtained dual image crack imagery exploitation weighted mean method
Red R, green G, the component of indigo plant B triple channels of picture are weighted average computation using following formula and obtain gray-scale map；
F (x, y)=0.299R (x, y)+0.587G (x, y)+0.114B (x, y) (1)
In above formula, f (x, y) is the gray value of image slices vegetarian refreshments, and R (x, y), G (x, y), B (x, y) are respectively color image pixel
Red R, green G, the component value of indigo plant B triple channels of point.
4. a kind of bridge bottom crack detection method based on binocular vision according to claim 3, it is characterised in that in
Value filtering denoising：Pending pixel, is placed on window center, obtains filter window by the filter window that make use of size to be 3*3
Comprising all pixels point gray value maximum and minimum value, if judge central pixel point gray value be equal to gray scale most
Big value or minimum value, then it is assumed that this pixel is noise spot, then is replaced using the gray scale intermediate value in window, if center pixel
The gray value of point is not equal to gray scale maximum or minimum value, then it is assumed that this pixel is signaling point, and the gray value of itself is as defeated
Go out.
5. a kind of bridge bottom crack detection method based on binocular vision according to claim 4, it is characterised in that figure
Image intensifying：The crack picture of above-mentioned medium filtering denoising is utilized into the piecewise linear function that threshold value is selected as given by following formula
Number carries out the brightness of specific tonal range in image enhancement processing, prominent image, and the color of crack and background is distinguished significantly；
<mrow>
<mi>G</mi>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mrow>
<mi>g</mi>
<mo>=</mo>
<msub>
<mi>g</mi>
<mi>min</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mfrac>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>-</mo>
<msub>
<mi>g</mi>
<mi>min</mi>
</msub>
<mo>)</mo>
<mo>&times;</mo>
<mn>255</mn>
</mrow>
<mrow>
<mo>(</mo>
<msub>
<mi>g</mi>
<mi>max</mi>
</msub>
<mo>-</mo>
<msub>
<mi>g</mi>
<mi>min</mi>
</msub>
<mo>)</mo>
</mrow>
</mfrac>
</mtd>
<mtd>
<mrow>
<msub>
<mi>g</mi>
<mi>min</mi>
</msub>
<mo><</mo>
<mi>g</mi>
<mo><</mo>
<msub>
<mi>g</mi>
<mi>max</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>255</mn>
</mtd>
<mtd>
<mrow>
<mi>g</mi>
<mo>=</mo>
<msub>
<mi>g</mi>
<mi>max</mi>
</msub>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
In above formula, G represents the gray value after pixel conversion, gmax、gminRepresent pixel in entire image gray scale maximum and
Minimum value, g represents the gray value before pixel conversion；With reference to enhancing effect, selected threshold is gmax=0.3, gmin=0.7.
6. a kind of bridge bottom crack detection method based on binocular vision according to claim 5, it is characterised in that
Sobel operators carry out edge extracting：Above-mentioned crack picture is subjected to rim detection using Sobel operators, will be each in image
The gray value f (x, y) of individual pixel carries out convolution algorithm with the two convolution kernels respectively, maximum is taken as output, by fortune
What is obtained after calculating is the image that a width embodies edge amplitude.
7. a kind of bridge bottom crack detection method based on binocular vision according to claim 2, it is characterised in that step
It is rapid 3) in, binocular camera is demarcated by Zhang Zhengyou standardizations, using the indeformable description operators of Fourier-Mellin
The matching of crack binary map obtained by binocular camera is carried out, world coordinate system is converted to finally by by the point under pixel coordinate system
Under point, the length and width in crack has been calculated using Euclidean distance formula.
8. a kind of bridge bottom crack detection method based on binocular vision according to claim 7, it is characterised in that double
Lens camera is demarcated：Three dimensional space coordinate is as follows relative to the corresponding relation of two-dimensional coordinate：
<mrow>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>x</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>y</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mi>a</mi>
<mi>M</mi>
<mi>V</mi>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>X</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Y</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Z</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mi>a</mi>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msub>
<mi>l</mi>
<mi>x</mi>
</msub>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<msub>
<mi>e</mi>
<mi>x</mi>
</msub>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<msub>
<mi>l</mi>
<mi>y</mi>
</msub>
</mtd>
<mtd>
<msub>
<mi>e</mi>
<mi>y</mi>
</msub>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>U</mi>
</mtd>
<mtd>
<mi>c</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>X</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Y</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Z</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
In above formula, M is the internal reference matrix of binocular camera, wherein lx、lyIt is binocular camera in both horizontally and vertically (x and y
Direction) on unit pixel size, ey、eyFor distortion factor of the binocular camera on both horizontally and vertically (x and y directions),
A is scale factor, and V is mapping matrix of the binocular camera coordinate system relative to world coordinate system, and wherein U is 3 rank spin matrixs,
C is 3*1 translation vector, and equation high order end is two-dimensional coordinate, and low order end is three dimensional space coordinate；What Zhang Zhengyou standardizations were carried out
It is plane reference, therefore makes scaling board be in Z=0 plane, then above formula is changed into：
<mrow>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>x</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>y</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mi>a</mi>
<mi>M</mi>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msub>
<mi>u</mi>
<mn>1</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>u</mi>
<mn>2</mn>
</msub>
</mtd>
<mtd>
<mi>c</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>X</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Y</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Z</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msub>
<mi>h</mi>
<mn>1</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>2</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>3</mn>
</msub>
</mtd>
</mtr>
</mtable>
</mfenced>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>X</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Y</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Z</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mi>H</mi>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>X</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Y</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>Z</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, the mapping matrix of three dimensional space coordinate to two-dimensional coordinate is H；
Position of the darkened features point under pixel coordinate system is recognized and calculated by image procossing and obtained in scaling board, and its world
Coordinate can be obtained by scaling board, as available from the above equation：
<mrow>
<mi>H</mi>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msub>
<mi>h</mi>
<mn>11</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>12</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>13</mn>
</msub>
</mtd>
</mtr>
<mtr>
<mtd>
<msub>
<mi>h</mi>
<mn>21</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>22</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>23</mn>
</msub>
</mtd>
</mtr>
<mtr>
<mtd>
<msub>
<mi>h</mi>
<mn>31</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>32</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>33</mn>
</msub>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msub>
<mi>h</mi>
<mn>1</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>2</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>h</mi>
<mn>3</mn>
</msub>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mi>a</mi>
<mi>M</mi>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msub>
<mi>u</mi>
<mn>1</mn>
</msub>
</mtd>
<mtd>
<msub>
<mi>u</mi>
<mn>2</mn>
</msub>
</mtd>
<mtd>
<mi>c</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
And then obtain：
Spin matrix U is unit orthogonal matrix, therefore column vector is orthogonal and is unit vector, then has：
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>u</mi>
<mn>1</mn>
<mi>T</mi>
</msubsup>
<msub>
<mi>u</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>u</mi>
<mn>1</mn>
</msub>
<mo>|</mo>
<mo>|</mo>
<mo>=</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>u</mi>
<mn>2</mn>
</msub>
<mo>|</mo>
<mo>|</mo>
<mo>=</mo>
<mn>1</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
</mrow>
Therefore all there are following restriction relations to the internal reference of calibration for cameras for any piece image, simultaneous formula (6), (7) can
：
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>h</mi>
<mn>1</mn>
<mi>T</mi>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>M</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<msup>
<mi>M</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msub>
<mi>h</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>h</mi>
<mn>1</mn>
<mi>T</mi>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>M</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<msup>
<mi>M</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msub>
<mi>h</mi>
<mn>1</mn>
</msub>
<mo>=</mo>
<msubsup>
<mi>h</mi>
<mn>2</mn>
<mi>T</mi>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>M</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<msup>
<mi>M</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msub>
<mi>h</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<mn>1</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
</mrow>
2
Projection matrix H is drawn by above formula.
9. a kind of bridge bottom crack detection method based on binocular vision according to claim 8, it is characterised in that figure
As matching：Two width binary maps of above-mentioned treated binocular camera are subjected to images match, it is assumed that two images f1(x, y) and
f2(x, y) has rotation, scaling, the relation of translation, then the relation between them can be expressed as follows：
f2(x, y)=f1(a(xcosβ+ysinβ)-Δx,a(-xsinβ+ycosβ)-Δy) (9)
Wherein, Δx、ΔyFor the translation vector on horizontally and vertically (x and y), a be proportional zoom between two width figures because
Son, is the anglec of rotation between two width figures；Carrying out Fourier transformation can obtain：
Wherein, f2The spectrum phase of (x, y) isIt is relevant with the anglec of rotation, translational movement and zoom factor, to (10) formula
Mould obtains power spectrum：
|F2(m, n) |=a-2|F1(a-1(mcosβ+nsinβ),a-1(-msinβ+ncosβ))| (11)
From (10) formula and (11) formula, image has zoom factor a, then its power spectrum just has zoom factor a-1, image is revolved
Turn, its anglec of rotation is β, then its power spectrum will rotate identical angle, and for spectrum center (m=n=0), it is to the anglec of rotation
Degree and yardstick are all constant, due to Δx、ΔyZoom factor and the anglec of rotation are determined, then a, β are transformed to translation form, first
Frequency spectrum is converted into polar coordinates：
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msup>
<mi>a</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mi>cos</mi>
<mi>&beta;</mi>
<mo>+</mo>
<mi>n</mi>
<mi>sin</mi>
<mi>&beta;</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mi>&rho;</mi>
<mi>a</mi>
</mfrac>
<mi>c</mi>
<mi>o</mi>
<mi>s</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&beta;</mi>
<mn>1</mn>
</msub>
<mo>-</mo>
<mi>&beta;</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msup>
<mi>a</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mi>m</mi>
<mi>sin</mi>
<mi>&beta;</mi>
<mo>+</mo>
<mi>n</mi>
<mi>c</mi>
<mi>o</mi>
<mi>s</mi>
<mi>&beta;</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mi>&rho;</mi>
<mi>a</mi>
</mfrac>
<mi>sin</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&beta;</mi>
<mn>1</mn>
</msub>
<mo>-</mo>
<mi>&beta;</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>12</mn>
<mo>)</mo>
</mrow>
</mrow>
Make hypothesis below：
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>S</mi>
<mi>&rho;</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>&beta;</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<mi>&rho;</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>|</mo>
<msub>
<mi>F</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>&rho;cos&beta;</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>&rho;sin&beta;</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>|</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>R</mi>
<mi>&rho;</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>&beta;</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<mi>&rho;</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>|</mo>
<msub>
<mi>F</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>&rho;cos&beta;</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>&rho;sin&beta;</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>|</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>13</mn>
<mo>)</mo>
</mrow>
</mrow>
Then by (12) Shi Ke get：
Sρ(β1, ρ) and=a-2Rρ((β1-β),ρ/a) (14)
λ=log ρ, k=loga are set simultaneously, then (14) formula can be deformed into：
Sρl(β1, ρ) and=a-2Rρl((β1-β),λ-k) (15)
(15) in formula, SρlRepresent logarithmic transformation, RρlIt is the indeformable description operators of Fourier-Mellin；It is changed into after conversion：
Sρl(m, n)=a-2Rρl(m,n)exp(-2jπ(mβ+nk)) (16)
Thus (16) formula can be converted to the gap of the anglec of rotation of two images and zoom factor the gap of translational movement, for flat
The gap of shifting amount, is tried to achieve using the Fourier inversion of crosspower spectrum.
10. a kind of bridge bottom crack detection method based on binocular vision according to claim 9, it is characterised in that
The calculating of fracture width and length：Obtained by images match after one group of pixel matched, i.e., binocular camera is clapped
The pixel taken the photograph in obtained crack picture, the projection matrix drawn is demarcated using binocular camera, just can draw what is matched
Pixel corresponding true point in three dimensions, two coordinates truly put imputed out are respectively (X, Y, Z), (P, Q, R),
<mrow>
<mi>L</mi>
<mo>=</mo>
<mroot>
<mrow>
<msup>
<mrow>
<mo>(</mo>
<mi>X</mi>
<mo>-</mo>
<mi>P</mi>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<mi>Y</mi>
<mo>-</mo>
<mi>Q</mi>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<mi>Z</mi>
<mo>-</mo>
<mi>R</mi>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
<mn>2</mn>
</mroot>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>17</mn>
<mo>)</mo>
</mrow>
</mrow>
The length and width in crack is then just can obtain using above formula distance between two points formula.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710452913.0A CN107179322A (en) | 2017-06-15 | 2017-06-15 | A kind of bridge bottom crack detection method based on binocular vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710452913.0A CN107179322A (en) | 2017-06-15 | 2017-06-15 | A kind of bridge bottom crack detection method based on binocular vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107179322A true CN107179322A (en) | 2017-09-19 |
Family
ID=59835649
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710452913.0A Pending CN107179322A (en) | 2017-06-15 | 2017-06-15 | A kind of bridge bottom crack detection method based on binocular vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107179322A (en) |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107545587A (en) * | 2017-09-26 | 2018-01-05 | 河北科技大学 | Round steel end face binocular visual positioning method based on major-minor eye |
| CN108647593A (en) * | 2018-04-26 | 2018-10-12 | 东华大学 | Unmanned plane road surface breakage classification and Detection method based on image procossing and SVM |
| CN108760767A (en) * | 2018-05-31 | 2018-11-06 | 电子科技大学 | Large-size LCD Screen defect inspection method based on machine vision |
| CN109186902A (en) * | 2018-09-26 | 2019-01-11 | 中国计量大学 | A kind of bridge structure health detection system of view-based access control model sensing |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN109754368A (en) * | 2019-01-23 | 2019-05-14 | 郑州工程技术学院 | A kind of crack joining method in bridge quality testing |
| CN109813509A (en) * | 2019-01-14 | 2019-05-28 | 中山大学 | The method that high-speed rail bridge vertically moves degree of disturbing measurement is realized based on unmanned plane |
| CN110009610A (en) * | 2019-03-27 | 2019-07-12 | 仲恺农业工程学院 | Visual detection method for surface damage of reservoir dam protection slope and bionic device |
| CN110223332A (en) * | 2019-06-10 | 2019-09-10 | 长安大学 | A kind of Bridge Crack scaling method |
| CN112051160A (en) * | 2020-09-09 | 2020-12-08 | 中山大学 | Segment joint bending stiffness measuring method, system, equipment and storage medium |
| CN112364802A (en) * | 2020-11-19 | 2021-02-12 | 中国地质调查局水文地质环境地质调查中心 | Deformation monitoring method for collapse landslide disaster body |
| CN112529880A (en) * | 2020-12-16 | 2021-03-19 | 武汉巨合科技有限公司 | Dam image crack detection device and method based on binocular vision |
| CN112630223A (en) * | 2020-12-07 | 2021-04-09 | 杭州申昊科技股份有限公司 | Tunnel-based crack detection system and method |
| CN112686877A (en) * | 2021-01-05 | 2021-04-20 | 同济大学 | Binocular camera-based three-dimensional house damage model construction and measurement method and system |
| CN113218838A (en) * | 2021-04-23 | 2021-08-06 | 江西理工大学 | Coal core permeability tester and testing method |
| CN113240635A (en) * | 2021-05-08 | 2021-08-10 | 中南大学 | Structural object detection image quality testing method with crack resolution as reference |
| CN114111602A (en) * | 2021-11-22 | 2022-03-01 | 招商局重庆交通科研设计院有限公司 | Bridge surface crack width calculation method based on image technology |
| CN114224489A (en) * | 2021-12-12 | 2022-03-25 | 浙江德尚韵兴医疗科技有限公司 | Trajectory tracking system for surgical robot and tracking method using the same |
| CN114638805A (en) * | 2022-03-21 | 2022-06-17 | 武汉纵横天地空间信息技术有限公司 | Track slab crack detection method and system and storage medium |
| CN114719749A (en) * | 2022-04-06 | 2022-07-08 | 重庆大学 | Metal surface crack detection and real size measurement method and system based on machine vision |
| CN117032276A (en) * | 2023-07-04 | 2023-11-10 | 长沙理工大学 | Bridge detection method and system based on binocular vision and inertial navigation fusion unmanned aerial vehicle |
| CN117825408A (en) * | 2024-03-05 | 2024-04-05 | 北京中科蓝图科技有限公司 | Integrated detection method, device and equipment for road |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN204142684U (en) * | 2014-10-10 | 2015-02-04 | 南京信息工程大学 | A kind of rotor wing unmanned aerial vehicle for wall hollow drum quality testing |
| CN104933718A (en) * | 2015-06-23 | 2015-09-23 | 广东省自动化研究所 | Physical coordinate positioning method based on binocular vision |
-
2017
- 2017-06-15 CN CN201710452913.0A patent/CN107179322A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN204142684U (en) * | 2014-10-10 | 2015-02-04 | 南京信息工程大学 | A kind of rotor wing unmanned aerial vehicle for wall hollow drum quality testing |
| CN104933718A (en) * | 2015-06-23 | 2015-09-23 | 广东省自动化研究所 | Physical coordinate positioning method based on binocular vision |
Non-Patent Citations (7)
| Title |
|---|
| 刘建威等: "《基于双目视觉的桥面裂缝检测技术研究》", 《铁道科学与工程学报》 * |
| 王琳等: "《基于双目立体视觉技术的桥梁裂缝测量***》", 《计算机应用》 * |
| 田浩等: "《数字图像处理、原理与应用》", 31 July 2007 * |
| 赵小川: "《现代数字图像处理技术提高及应用案例详解MATLAB版》", 30 April 2012 * |
| 钟志光等: "《Visual C++，NET数字图像处理实例与解析》", 30 June 2003 * |
| 陈爱军等: "《数字图像处理及其MATLAB实现》", 31 July 2008 * |
| 鱼滨等: "《基于MATLAB和遗传算法的图像处理》", 31 October 2015 * |
Cited By (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107545587A (en) * | 2017-09-26 | 2018-01-05 | 河北科技大学 | Round steel end face binocular visual positioning method based on major-minor eye |
| CN108647593A (en) * | 2018-04-26 | 2018-10-12 | 东华大学 | Unmanned plane road surface breakage classification and Detection method based on image procossing and SVM |
| CN108760767A (en) * | 2018-05-31 | 2018-11-06 | 电子科技大学 | Large-size LCD Screen defect inspection method based on machine vision |
| CN108760767B (en) * | 2018-05-31 | 2020-07-17 | 电子科技大学 | Large-size liquid crystal display defect detection method based on machine vision |
| CN109186902A (en) * | 2018-09-26 | 2019-01-11 | 中国计量大学 | A kind of bridge structure health detection system of view-based access control model sensing |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN109813509A (en) * | 2019-01-14 | 2019-05-28 | 中山大学 | The method that high-speed rail bridge vertically moves degree of disturbing measurement is realized based on unmanned plane |
| CN109813509B (en) * | 2019-01-14 | 2020-01-24 | 中山大学 | Method for realizing measurement of vertical dynamic disturbance degree of high-speed rail bridge based on unmanned aerial vehicle |
| CN109754368A (en) * | 2019-01-23 | 2019-05-14 | 郑州工程技术学院 | A kind of crack joining method in bridge quality testing |
| CN110009610A (en) * | 2019-03-27 | 2019-07-12 | 仲恺农业工程学院 | Visual detection method for surface damage of reservoir dam protection slope and bionic device |
| CN110223332A (en) * | 2019-06-10 | 2019-09-10 | 长安大学 | A kind of Bridge Crack scaling method |
| CN110223332B (en) * | 2019-06-10 | 2021-08-31 | 长安大学 | Bridge crack calibration method |
| CN112051160A (en) * | 2020-09-09 | 2020-12-08 | 中山大学 | Segment joint bending stiffness measuring method, system, equipment and storage medium |
| CN112051160B (en) * | 2020-09-09 | 2022-04-19 | 中山大学 | Segment joint bending stiffness measuring method, system, equipment and storage medium |
| CN112364802B (en) * | 2020-11-19 | 2021-08-03 | 中国地质调查局水文地质环境地质调查中心 | Deformation monitoring method for collapse landslide disaster body |
| CN112364802A (en) * | 2020-11-19 | 2021-02-12 | 中国地质调查局水文地质环境地质调查中心 | Deformation monitoring method for collapse landslide disaster body |
| CN112630223A (en) * | 2020-12-07 | 2021-04-09 | 杭州申昊科技股份有限公司 | Tunnel-based crack detection system and method |
| CN112630223B (en) * | 2020-12-07 | 2023-12-26 | 杭州申昊科技股份有限公司 | Tunnel crack detection system and method |
| CN112529880A (en) * | 2020-12-16 | 2021-03-19 | 武汉巨合科技有限公司 | Dam image crack detection device and method based on binocular vision |
| CN112686877B (en) * | 2021-01-05 | 2022-11-11 | 同济大学 | Binocular camera-based three-dimensional house damage model construction and measurement method and system |
| CN112686877A (en) * | 2021-01-05 | 2021-04-20 | 同济大学 | Binocular camera-based three-dimensional house damage model construction and measurement method and system |
| CN113218838A (en) * | 2021-04-23 | 2021-08-06 | 江西理工大学 | Coal core permeability tester and testing method |
| CN113240635B (en) * | 2021-05-08 | 2022-05-03 | 中南大学 | Structural object detection image quality testing method with crack resolution as reference |
| CN113240635A (en) * | 2021-05-08 | 2021-08-10 | 中南大学 | Structural object detection image quality testing method with crack resolution as reference |
| CN114111602A (en) * | 2021-11-22 | 2022-03-01 | 招商局重庆交通科研设计院有限公司 | Bridge surface crack width calculation method based on image technology |
| CN114111602B (en) * | 2021-11-22 | 2023-07-25 | 招商局重庆交通科研设计院有限公司 | Bridge surface crack width calculation method based on image technology |
| CN114224489A (en) * | 2021-12-12 | 2022-03-25 | 浙江德尚韵兴医疗科技有限公司 | Trajectory tracking system for surgical robot and tracking method using the same |
| CN114224489B (en) * | 2021-12-12 | 2024-02-13 | 浙江德尚韵兴医疗科技有限公司 | Track tracking system for surgical robot and tracking method using same |
| CN114638805A (en) * | 2022-03-21 | 2022-06-17 | 武汉纵横天地空间信息技术有限公司 | Track slab crack detection method and system and storage medium |
| CN114719749A (en) * | 2022-04-06 | 2022-07-08 | 重庆大学 | Metal surface crack detection and real size measurement method and system based on machine vision |
| CN114719749B (en) * | 2022-04-06 | 2023-07-14 | 重庆大学 | Metal surface crack detection and real size measurement method and system based on machine vision |
| CN117032276A (en) * | 2023-07-04 | 2023-11-10 | 长沙理工大学 | Bridge detection method and system based on binocular vision and inertial navigation fusion unmanned aerial vehicle |
| CN117032276B (en) * | 2023-07-04 | 2024-06-25 | 长沙理工大学 | Bridge detection method and system based on binocular vision and inertial navigation fusion unmanned aerial vehicle |
| CN117825408A (en) * | 2024-03-05 | 2024-04-05 | 北京中科蓝图科技有限公司 | Integrated detection method, device and equipment for road |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107179322A (en) | A kind of bridge bottom crack detection method based on binocular vision | |
| CN105279372B (en) | A kind of method and apparatus of determining depth of building | |
| CN110569704B (en) | Multi-strategy self-adaptive lane line detection method based on stereoscopic vision | |
| Lei et al. | New crack detection method for bridge inspection using UAV incorporating image processing | |
| CN109615611B (en) | Inspection image-based insulator self-explosion defect detection method | |
| CN104574393B (en) | A kind of three-dimensional pavement crack pattern picture generates system and method | |
| CN102297660B (en) | Measuring method of shield tunnel lining segment seam open width and apparatus thereof | |
| CN112733950A (en) | Power equipment fault diagnosis method based on combination of image fusion and target detection | |
| CN108416771A (en) | Metal material corrosion area detection method based on monocular camera | |
| CN109410207A (en) | A kind of unmanned plane line walking image transmission line faultlocating method based on NCC feature | |
| CN108195736B (en) | Method for extracting vegetation canopy clearance rate through three-dimensional laser point cloud | |
| CN108665468B (en) | Device and method for extracting tangent tower insulator string | |
| CN104700395A (en) | Method and system for detecting appearance crack of structure | |
| CN115272271A (en) | Pipeline defect detecting and positioning ranging system based on binocular stereo vision | |
| CN109919856B (en) | Asphalt pavement structure depth detection method based on binocular vision | |
| CN109492525B (en) | Method for measuring engineering parameters of base station antenna | |
| CN110189375A (en) | A kind of images steganalysis method based on monocular vision measurement | |
| CN115830004A (en) | Surface defect detection method, device, computer equipment and storage medium | |
| CN107590444A (en) | Detection method, device and the storage medium of static-obstacle thing | |
| CN107462182A (en) | A kind of cross section profile deformation detecting method based on machine vision and red line laser | |
| CN116152697A (en) | Three-dimensional model measuring method and related device for concrete structure cracks | |
| CN114842340A (en) | Robot binocular stereoscopic vision obstacle sensing method and system | |
| CN105115443B (en) | The full visual angle high precision three-dimensional measurement method of level of view-based access control model e measurement technology | |
| CN112613437B (en) | Method for identifying illegal buildings | |
| CN112801141B (en) | Heterogeneous image matching method based on template matching and twin neural network optimization |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170919 |