CN109523528A - A kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm - Google Patents
A kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm Download PDFInfo
- Publication number
- CN109523528A CN109523528A CN201811340192.5A CN201811340192A CN109523528A CN 109523528 A CN109523528 A CN 109523528A CN 201811340192 A CN201811340192 A CN 201811340192A CN 109523528 A CN109523528 A CN 109523528A
- Authority
- CN
- China
- Prior art keywords
- algorithm
- pixel
- transmission line
- electricity
- image
- 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.)
- Granted
Links
- 230000005540 biological transmission Effects 0.000 title claims abstract description 71
- 230000005611 electricity Effects 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000012937 correction Methods 0.000 claims abstract description 24
- 230000011218 segmentation Effects 0.000 claims abstract description 20
- 230000002902 bimodal effect Effects 0.000 claims abstract description 9
- 238000003384 imaging method Methods 0.000 claims abstract description 8
- 238000010276 construction Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 238000009499 grossing Methods 0.000 claims description 5
- 230000003190 augmentative effect Effects 0.000 claims description 3
- 238000004886 process control Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 3
- 238000007689 inspection Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 9
- 230000003278 mimic effect Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000001027 hydrothermal synthesis Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 239000012212 insulator Substances 0.000 description 1
- 238000009533 lab test Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000002023 wood Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
- G06T7/85—Stereo camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/80—Geometric correction
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of transmission line of electricity extracting methods based on unmanned plane binocular vision SGC algorithm, comprising steps of Image Acquisition, the acquisition of image information is the basis of binocular vision system, Image Acquisition is influenced by natural environment, camera properties, shooting level, and the image data of high quality will provide great convenience for the image procossing in later period；Binocular camera demarcates the inside and outside parameter for needing to obtain left and right camera respectively, determines the relative positional relationship of two cameras, finally establishes binocular imaging model, instructs the calculating of target three-dimensional information；Image rectification is collecting initial pictures to, needing to carry out color correction and geometric correction to it later；It corrects image and carries out SGC Stereo matching, obtain disparity map；The analysis for completing grey level histogram is carried out the segmentation of disparity map using bimodal Threshold Segmentation Algorithm, obtains pure transmission line of electricity.Using technical solution of the present invention, the detection accuracy of transmission line of electricity can be largely improved.
Description
Technical field
The invention belongs to the inspection of power transmission line unmanned machine and technical field of computer vision, and in particular to one kind is based on nobody
The transmission line of electricity extracting method of machine binocular vision SGC algorithm.
Background technique
China's power demands are larger, and transmission line of electricity is widely distributed, wherein be no lack of across dangerously steep mountain area and complicated landform, it is right
The defects of in the transmission line of electricity for being in dangerously steep mountain area, there are some hidden danger to be difficult to find, such as wire strand breakage, foreign matter are hung.It adopts
Power circuit polling is carried out with unmanned plane, it is not only time saving and energy saving, it is more suitable for the inspection demand of the transmission line of electricity in dangerously steep area.Electricity
Unmanned plane inspection has been applied to the inspection of transmission line of electricity by net company, and establishes the inspection that unmanned plane is combined with manual inspection
Mode.
Currently, domestic and foreign scholars have done numerous studies for polling transmission line diagnostic imaging, but mostly both for biography
The transmission line of electricity two dimensional image of system carries out different image processing algorithm research, it is intended to improve the effect of transmission line of electricity defect recognition
Rate and accuracy rate.But since unmanned plane inspection transmission line of electricity image background is extremely complex, including mountains and rivers, river, road, tree
Wood, farmland and building etc., these complicated backgrounds directly affect traditional image recognition effect.If can be by unmanned plane
The transmission line of electricity and complex background of prospect distinguish in inspection picture, and the defect recognition for greatly improving transmission line of electricity is accurate
Rate.
Binocular vision and Depth Imaging are to realize the effective way of the demand, i.e. UAV flight's binocular imaging apparatus,
The binocular image for obtaining transmission line of electricity obtains the depth information of each target in image using Stereo Matching Algorithm, is based on depth
Information can separate prospect power transmission line with complex background.At present for binocular vision and depth imaging technique in transmission line of electricity
Inspection application aspect, has no relevant report.In terms of binocular vision research, there are higher level in foreign countries, and China is in the field
Research also have made some progress.Hrabar use binocular vision method and LiDAR equipment, can detecte large-scale shaft tower and
Trees, but cannot detect the position of conducting wire；University Of Chongqing Yang Hao et al. is using technique of binocular stereoscopic vision to -- icing insulator
It is detected；Central China University of Science and Technology Chang Wenkai et al. is believed using binocular stereo vision close-in measurement conducting wire partial 3 d position
Breath, measurement result and actual wire width are very close.In terms of solid matching method research, absolute error and algorithm (SAD,
Sum of Absolute Differences) it is used for image Block- matching, the absolute value of the difference that each pixel corresponds to numerical value is asked
With to assess the similarity of two image blocks, be usually used in the preliminary screening of multistep treatment；Roy earliest by figure cut algorithm (GC,
Graph Cuts) be applied to Stereo matching in, the experimental results showed that for global registration algorithm there are the shortcomings that, GC Stereo matching
Algorithm can effectively overcome, and can avoid parallax discontinuous problem near polar curve.
Summary of the invention
In order to solve the above technical problems in background technology, the present invention provides it is a kind of can will be complicated and changeable
Natural background and transmission line of electricity are carried out effectively segmentation and obtain being transmitted electricity based on unmanned plane binocular vision SGC for pure transmission line of electricity image
The method that route extracts, this method are directed to the problem of unmanned plane inspection complex background transmission line of electricity image recognition, vertical using SGC
Body matching algorithm and Threshold Segmentation Algorithm handle power transmission line binocular image.Wherein, SGC Stereo Matching Algorithm obtains transmission of electricity
The effective disparity map of line image is split disparity map using Threshold Segmentation Algorithm complex background rejecting the pure transmission of electricity of acquisition
Line can largely improve the detection accuracy of transmission line of electricity.
The present invention adopts the following technical scheme that realize:
A kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm, comprising the following steps:
(1) stereo calibration is carried out to left and right camera, obtains the inside and outside parameter of left and right camera respectively, determines two cameras
Relative positional relationship finally establishes binocular imaging model, instructs the calculating of target three-dimensional information；
(2) transmission line of electricity left images are obtained by UAV flight's binocular camera；
(3) figure of the initial pictures to color correction and geometric correction later, is carried out to it, after obtaining polar curve correction is collected
Picture；
(4) processing of SGC Stereo Matching Algorithm is carried out to the image after polar curve correction, obtains disparity map；
(5) segmentation that disparity map is carried out using bimodal Threshold Segmentation Algorithm, obtains pure transmission line of electricity.
A further improvement of the present invention lies in that carrying out stereo calibration to left and right camera in step (1), specific method is such as
Under:
By adjusting the direction of gridiron pattern scaling board, the photo of different directions is shot for gridiron pattern scaling board, is acquired respectively
Each 18 left images of left images, the calibrating procedure of direct OpenCV demarcate binocular camera.
A further improvement of the present invention lies in that binocular camera is mounted in automatic collection or so figure on unmanned plane in step (2)
Picture, the specific method is as follows:
By the process control binocular camera in the microcomputer that is carried on unmanned plane, and set between the time of acquisition
Every realization automatic collection or so transmission of electricity line image pair during unmanned plane during flying.
A further improvement of the present invention lies in that being carried out in step (3) to the left and right two width transmission of electricity line image of unmanned plane acquisition
Polar curve correction, the specific method is as follows:
Polar curve is corrected by image, to the image after being corrected after the progress binocular correction of binocular left and right image, from figure
In be able to observe that, in left and right image in corresponding pixel on the parallel polar curve of same.
A further improvement of the present invention lies in that carrying out SGC Stereo Matching Algorithm to the image after polar curve correction in step (4)
Processing, the specific method is as follows:
(401) construction of new energy function
Complete grid chart energy function consists of two parts: first part indicates the gray value of each pixel itself, the
Two parts then indicate the smoothing properties of neighbor pixel pair, so smooth side decides energy function in complete grid chart
Value, and in simplified grid chart, have by the calculated priori parallax value of sectional perspective matching algorithm SAD and with smooth side
The penalty term of pass codetermines energy function value；
When construction matches energy function, two bound terms of data item and smooth item are needed:
In formula:
F --- it is marked for parallax；
Edata(f) --- data item represents the inconsistency of data observed by f；
Esmooth(f) --- smooth item indicates the degree of f local smoothing method；
P --- the set of picture all pixels point；
Q a --- pixel in set；
fp--- the label parallax of p；
fq--- the label parallax of q；
{ p, q } --- two adjacent pixels in picture；
N --- adjacent pixel is to set in picture；
In data item, using the linear interpolation method of pixel dissmilarity degree, i.e., the dissimilar degree of two pixels is not
The dissimilarity of pixel grey scale can be only used to measure, therefore, using the method for sub-pix linear interpolation, single pixel is clicked through
Row matching, can obtain more accurately matching in this way, specific formula is as shown:
Dp(fp)=| I1(x,y)-I2(x+fp,y)| (2)
In formula (2), fp∈{f1,p,f2,p,…,fN,p, I1(x, y) represents the gray value of pixel；
Linear interpolation first is carried out to x-axis:
It expands in y-axis again, as follows:
In order to improve arithmetic speed, the possibility disparity range of pixel is first first obtained using SAD Local Optimization Algorithm, then carry out
It is matched to two-dimensional sub-pix；
In order to avoid generating excessive smooth phenomenon in edge and multi-texturing region, using the smooth function of Potts model,
The formula of smooth item are as follows:
Vp,q(fp,fq)=λ T (5)
In formula (5) and (6), wherein k is constant, Ip, IqIt is the grey scale pixel value at point p and q respectively, works as fp=fqWhen,
T=0；Work as fp≠fqWhen, T=1；
(402) simplify the building of grid chart
First with SAD Stereo Matching Algorithm, in parallax [dmin,dmax] each pixel in left view is calculated in range
Then matching value chooses N number of most suitable matching value, the pixel parallax selected is labeled as { d as initial estimate1,p,
d2,p,…,dN,p, corresponding matching value is
In order to simplify grid chart, each pixel in figure only retains N-1 node and N side, delete remaining node and
Edge；Therefore, the node around each pixel in simplified grid chart isBy parallax side by these nodes
It links together, the capacity of each edge isC is the constant for meeting formula (7), and the purpose for introducing constant C is to protect
Card t-links is cut just in a line circle；
The energy function of complete grid chart includes two parts, and first part represents pixel gray value itself；Second part
For the smooth attribute of adjacent pixel；Therefore, in complete grid chart, the value of energy function depends on the side n-links；Simplifying net
In trrellis diagram, the two parts that have of energy function value are determined, a part is using the calculated priori parallax value of local matching, another portion
Dividing is penalty term related with the side n-links；
Shown in simplified grid chart energy function formula such as formula (8):
In formula (8), fpParallax label is represented, simplified parallax tag set only includes N number of potential possible parallax value；
Dp(fp) indicate parallax label fpIt is assigned to the cost function of pixel p；u{p,q}|fP-fq| indicate the smooth letter of Potts model
Number；
(403) matching that energy function minimizes
After the grid chart construction complete for reducing optimization, carry out s-t cutting, by vertex V be divided into two it is non-intersecting
Set S and set T, in this way, be equivalent to find the cutting C of minimum cost, i.e. C=S, T the problem of minimal cut, further according to
The augmenting path algorithm of Ford and Fulkerson is calculated in the minimum value for seeking energy function using max-flow/minimal cut
Method, label f corresponding to minimal cut, meets E (f)=argminE (f), and reusing α-expansion algorithm, to carry out energy function minimum
Change and solves, the distribution of parallax required for finally obtaining.
A further improvement of the present invention lies in that k value is 20.
A further improvement of the present invention lies in that disparity map is split using bimodal Threshold Segmentation Algorithm in step (5),
Removal complex background obtains pure defeated circuit image, and the specific method is as follows:
It is analyzed according to the grey level histogram to transmission line of electricity disparity map, is split using bimodal Threshold Segmentation Algorithm
Rejecting complex background is split to disparity map and obtains pure transmission line of electricity image.
The present invention has following beneficial technical effect:
1, a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm provided by the invention, Neng Gouzhun
It really detects transmission line of electricity, and complex background is rejected and obtains pure transmission line of electricity image, be the fault diagnosis of next step
It provides convenience, improves the accuracy of transmission line malfunction detection.
2, a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm provided by the invention, avoids
Three-dimensional reconstruction is carried out to transmission line of electricity, the relative depth information only calculated between target object completes image segmentation, greatly reduces
Computation complexity can be good at the requirement of real-time for meeting transmission line faultlocating.
Detailed description of the invention
Fig. 1 is a kind of transmission line of electricity extracting method flow chart based on unmanned plane binocular vision SGC algorithm of the invention；
Fig. 2 is the flow chart of SGC algorithm Stereo Matching Algorithm of the invention；
Fig. 3 is the simplified grid chart of the present invention；
Fig. 4 is standard testing picture of the present invention；
Fig. 5 is algorithms of different disparity map of the present invention comparison, wherein (a) is true disparity map, it is (b) SAD algorithm process institute
, (c) obtained by GC algorithm process, (d) obtained by SGC algorithm process；
Fig. 6 is laboratory test results of the present invention, wherein (a) is left image, (b) is right image, (c) is disparity map；
Fig. 7 is simulated experiment scene of the present invention, wherein (a) is left image, (b) is right image, (c) is disparity map；
Fig. 8 is actual scene of the present invention, wherein (a) is left image, (b) is right image, (c) is disparity map；
Fig. 9 is grey level histogram of the invention, wherein (a) is the grey level histogram of mimic transmission line scene one, it (b) is mould
The grey level histogram of quasi- power transmission line scene two is (c) grey level histogram of mimic transmission line scene three, (d) is practical power transmission line
Grey level histogram；
Figure 10 is the segmentation figure of power transmission line of the present invention, wherein (a) is the segmentation figure of mimic transmission line scene one, it (b) is mould
The segmentation figure of quasi- power transmission line scene two is (c) segmentation figure of mimic transmission line scene three, (d) is the segmentation figure of practical power transmission line；
Specific embodiment
For the purpose of the present invention, technical solution and feature is more clearly understood, below in conjunction with reference attached drawing, to the present invention
It is further described.
As shown in Figure 1, a kind of transmission line of electricity extraction side based on unmanned plane binocular vision SGC algorithm provided by the invention
Method, comprising the following steps:
(1) binocular camera calibration needs to obtain the inside and outside parameter of left and right camera respectively, determines the relative position of two cameras
Relationship finally establishes binocular imaging model, instructs the calculating of target three-dimensional information；
(2) by binocular camera and microcomputer-equipped on unmanned plane, secondary open is carried out by the SDK provided camera
Hair realizes binocular camera automatic collection or so transmission of electricity line image pair；
(3) stereo calibration directly is carried out to binocular camera using the calibrating procedure of OpenCV (open source computer vision library)；
The binocular camera of UAV flight collects left images to, needing to carry out color correction and geometric correction to it later, obtains
Picture after polar curve correction；
(4) using the transmission line of electricity image pair of the left and right camera shooting after the processing correction of SGC Stereo Matching Algorithm, had
The disparity map of effect；
(5) grey level histogram for analyzing disparity map, is split disparity map using bimodal Threshold Segmentation Algorithm, rejects figure
Complex background as in obtains pure transmission line of electricity image.
Calibration tool demarcates binocular camera using OpenCV, and specific demarcation flow is as follows:
By adjusting the direction of gridiron pattern scaling board, the photo of different directions is shot for gridiron pattern scaling board, is acquired respectively
Each 18 left images of left images；
Binocular camera demarcates the inside and outside parameter for needing to obtain left and right camera respectively, determines that the relative position of two cameras is closed
System, finally establishes binocular imaging model, instructs the calculating of target three-dimensional information.
By the process control binocular camera in the microcomputer that is carried on unmanned plane, and set between the time of acquisition
Every realization automatic collection or so transmission line of electricity image pair during unmanned plane during flying.
The inside and outside ginseng obtained using calibration, to polar curve correction is carried out, needs to carry out color to it to collected left images
Correction and geometric correction, the picture after obtaining polar curve correction.
It is as follows that SGC matches Stereo Matching Algorithm process:
The construction of new energy function
Complete grid chart energy function consists of two parts: first part indicates the gray value of each pixel itself, the
Two parts then indicate the smoothing properties of neighbor pixel pair.So smooth side decides energy function in complete grid chart
Value.And in simplified grid chart, have by the calculated priori parallax value of sectional perspective matching algorithm SAD and with smooth side
The penalty term of pass codetermines energy function value.
When construction matches energy function, two bound terms of data item and smooth item are needed.
In data item, using the linear interpolation method of pixel dissmilarity degree, i.e., the dissimilar degree of two pixels is not
The dissimilarity of pixel grey scale can be only used to measure.Therefore, using the method for sub-pix linear interpolation, single pixel is clicked through
Row matching, can obtain more accurately matching in this way, specific formula is as shown:
Dp(fp)=| I1(x,y)-I2(x+fp,y)|
Linear interpolation first is carried out to x-axis:
It expands in y-axis again:
In order to improve arithmetic speed, the possibility disparity range of pixel is first first obtained using SAD Local Optimization Algorithm, then carry out
It is matched to two-dimensional sub-pix.
In order to avoid generating excessive smooth phenomenon in edge and multi-texturing region, using the smooth function of Potts model,
The formula of smooth item are as follows:
Vp,q(fp,fq)=λ T
Ip, IqIt is the grey scale pixel value at point p and q respectively.K=20 herein, works as fp=fqWhen, T=0；Work as fp≠fqWhen,
T=1.
Simplify the building of grid chart
The present invention constructs a simplified grid chart, deletes some nodes and edge, so that the scale of figure is substantially reduced,
For each pixel of image, only retain some possible parallax values, therefore does not need to traverse all disparity ranges, it can be with
Greatly improve arithmetic speed.
First with SAD Stereo Matching Algorithm, in parallax [dmin,dmax] each pixel in left view is calculated in range
Then matching value chooses N number of most suitable matching value.Theoretically speaking N value is bigger, matched effect is better, but matches
Time is longer；N value is smaller, and match time is shorter, but matching effect is poorer.Therefore the value of N will be simultaneously in view of matching effect
Fruit and match time, according to gradient of disparity theory, the parallax of most of points is between 0~4, so not losing general in image
Property, under normal circumstances, select N=4.The pixel parallax selected is labeled as { d as initial estimate1,p,d2,p,…,dN,p, phase
The matching value answered is
In order to simplify grid chart, each pixel in figure only retains N-1 node and N side, delete remaining node and
Edge.Therefore, the node around each pixel in simplified grid chart isPass through t-links (connection pixel
To the side of terminal node) by these node contacts together.The capacity of each edge isC is a constant, is introduced normal
The purpose for measuring C is to guarantee that t-links is cut just in a line circle.
The energy function of complete grid chart includes two parts, and first part represents pixel gray value itself；Second part
For the smooth attribute of adjacent pixel.Therefore, in complete grid chart, the value of energy function depends on n-links (connection nonterminal
The side of node).In simplifying grid chart, the two parts that have of energy function value are determined, a part is calculated using local matching
Priori parallax value, another part is penalty term related with the side n-links.
The solid stain of attached drawing 2 indicates the potentially possible parallax value point retained during building grid chart, solid line then table
Show that simplified grid edge, hollow dots and dotted line then indicate the node and edge deleted.Therefore, simplified grid chart energy
Shown in function expression such as formula (8):
In formula (8), fpParallax label is represented, simplified parallax tag set only includes N number of potential possible parallax value；
Dp(fp) indicate parallax label fpIt is assigned to the cost function of pixel p；u{p,q}|fP-fq| indicate the smooth letter of Potts model
Number.
The matching that energy function minimizes
After the grid chart construction complete for reducing optimization, a s-t cutting (C=S, T) is carried out by vertex V and is divided into two
Disjoint set S and set T, in this way, being equivalent to find the cutting C of minimum cost the problem of minimal cut, further according to Ford
With the augmenting path algorithm of Fulkerson, in the minimum value for seeking energy function, using max-flow/minimal cut algorithm, most
It is small to cut corresponding label f, meet E (f)=argminE (f), reuses α-expansion algorithm progress energy function minimum and ask
Solution, finally obtains required disparity map.
According to the comparison for the treatment of effect shown in table 1 and time, the runing time of SAD algorithm can be significantly seen very
Fastly, but error hiding rate is very high, and there are many Mismatching point；The error hiding rate of traditional GC Stereo Matching Algorithm is very low, but the time
Complexity is very high, is unfavorable for common engineer application；The error hiding rate of this paper modified hydrothermal process (SGC) is close to traditional GC
Stereo Matching Algorithm, it is more many than the high treating effect of SAD, but runing time is traditional GC Stereo Matching Algorithm 1/10th
Left and right can be applied to Practical Project close to real-time processing.
Table 1
It is analyzed according to the grey level histogram to transmission line of electricity disparity map, is split using bimodal Threshold Segmentation Algorithm
Rejecting complex background is split to disparity map and obtains pure transmission line of electricity image.
Claims (7)
1. a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm, which is characterized in that including following step
It is rapid:
(1) stereo calibration is carried out to left and right camera, obtains the inside and outside parameter of left and right camera respectively, determines the opposite of two cameras
Positional relationship finally establishes binocular imaging model, instructs the calculating of target three-dimensional information；
(2) transmission line of electricity left images are obtained by UAV flight's binocular camera；
(3) image of the initial pictures to color correction and geometric correction later, is carried out to it, after obtaining polar curve correction is collected；
(4) processing of SGC Stereo Matching Algorithm is carried out to the image after polar curve correction, obtains disparity map；
(5) segmentation that disparity map is carried out using bimodal Threshold Segmentation Algorithm, obtains pure transmission line of electricity.
2. a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm according to claim 1, special
Sign is, carries out stereo calibration to left and right camera in step (1), the specific method is as follows:
By adjusting the direction of gridiron pattern scaling board, the photo of different directions is shot for gridiron pattern scaling board, respectively acquisition left and right
Each 18 left images of image, the calibrating procedure of direct OpenCV demarcate binocular camera.
3. a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm according to claim 2, special
Sign is that binocular camera is mounted in automatic collection left images on unmanned plane in step (2), and the specific method is as follows:
By the process control binocular camera in the microcomputer that carries on unmanned plane, and the time interval of acquisition is set,
Realize automatic collection or so transmission of electricity line image pair during unmanned plane during flying.
4. a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm according to claim 3, special
Sign is, carries out polar curve correction to the left and right two width transmission of electricity line image of unmanned plane acquisition in step (3), the specific method is as follows:
Polar curve is corrected by image, to the image after being corrected after the progress binocular correction of binocular left and right image, the energy from figure
It enough observes, in left and right image in corresponding pixel on the parallel polar curve of same.
5. a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm according to claim 4, special
Sign is, carries out the processing of SGC Stereo Matching Algorithm to the image after polar curve correction in step (4), the specific method is as follows:
(401) construction of new energy function
Complete grid chart energy function consists of two parts: first part indicates the gray value of each pixel itself, second
Divide the smoothing properties for then indicating neighbor pixel pair, so smooth side decides the value of energy function in complete grid chart, and
In simplified grid chart, by the calculated priori parallax value of sectional perspective matching algorithm SAD and related with smooth side
Penalty term codetermines energy function value；
When construction matches energy function, two bound terms of data item and smooth item are needed:
In formula:
F --- it is marked for parallax；
Edata(f) --- data item represents the inconsistency of data observed by f；
Esmooth(f) --- smooth item indicates the degree of f local smoothing method；
P --- the set of picture all pixels point；
Q a --- pixel in set；
fp--- the label parallax of p；
fq--- the label parallax of q；
{ p, q } --- two adjacent pixels in picture；
N --- adjacent pixel is to set in picture；
In data item, using the linear interpolation method of pixel dissmilarity degree, i.e., the dissimilar degree of two pixels cannot be only
It is measured using the dissimilarity of pixel grey scale, therefore, using the method for sub-pix linear interpolation, to the progress of single pixel point
Match, can obtain more accurately matching in this way, specific formula is as shown:
Dp(fp)=| I1(x,y)-I2(x+fp,y)| (2)
In formula (2), fp∈{f1,p,f2,p,…,fN,p, I1(x, y) represents the gray value of pixel；
Linear interpolation first is carried out to x-axis:
It expands in y-axis again, as follows:
In order to improve arithmetic speed, the possibility disparity range of pixel is first first obtained using SAD Local Optimization Algorithm, then proceed to two
The sub-pix of dimension matches；
In order to avoid generating excessive smooth phenomenon at edge and multi-texturing region, using the smooth function of Potts model, smoothly
The formula of item are as follows:
Vp,q(fp,fq)=λ T (5)
In formula (5) and (6), wherein k is constant, Ip, IqIt is the grey scale pixel value at point p and q respectively, works as fp=fqWhen, T=0；
Work as fp≠fqWhen, T=1；
(402) simplify the building of grid chart
First with SAD Stereo Matching Algorithm, in parallax [dmin,dmax] matching of each pixel in left view is calculated in range
Then value chooses N number of most suitable matching value, the pixel parallax selected is labeled as { d as initial estimate1,p,d2,p,…,
dN,p, corresponding matching value is
In order to simplify grid chart, each pixel in figure only retains N-1 node and N side, deletes remaining node and edge；
Therefore, the node around each pixel in simplified grid chart isBy parallax side by these node contacts
Together, the capacity of each edge isC is the constant for meeting formula (7), and the purpose for introducing constant C is to guarantee t-
Links is cut just in a line circle；
The energy function of complete grid chart includes two parts, and first part represents pixel gray value itself；Second part is phase
The smooth attribute of adjacent pixel；Therefore, in complete grid chart, the value of energy function depends on the side n-links；Simplifying grid chart
In, determine the two parts that have of energy function value, a part is using the calculated priori parallax value of local matching, and another part is
Penalty term related with the side n-links；
Shown in simplified grid chart energy function formula such as formula (8):
In formula (8), fpParallax label is represented, simplified parallax tag set only includes N number of potential possible parallax value；Dp
(fp) indicate parallax label fpIt is assigned to the cost function of pixel p；u{p,q}|fP-fq| indicate the smooth letter of Potts model
Number；
(403) matching that energy function minimizes
After the grid chart construction complete for reducing optimization, a s-t cutting is carried out, vertex V is divided into two disjoint sets
Close S and set T, in this way, be equivalent to find the cutting C of minimum cost, i.e. C=S, T the problem of minimal cut, further according to Ford with
The augmenting path algorithm of Fulkerson, in the minimum value for seeking energy function, using max-flow/minimal cut algorithm, minimal cut
Corresponding label f, meets E (f)=argminE (f), reuses α-expansion algorithm and carries out energy function minimum solution, most
The distribution of parallax required for obtaining eventually.
6. a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm according to claim 5, special
Sign is that k value is 20.
7. a kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm according to claim 5, special
Sign is, is split to disparity map using bimodal Threshold Segmentation Algorithm in step (5), and removal complex background obtains pure defeated
Circuit image, the specific method is as follows:
It is analyzed according to the grey level histogram to transmission line of electricity disparity map, is split using bimodal Threshold Segmentation Algorithm to view
Poor figure is split rejecting complex background and obtains pure transmission line of electricity image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811340192.5A CN109523528B (en) | 2018-11-12 | 2018-11-12 | Power transmission line extraction method based on unmanned aerial vehicle binocular vision SGC algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811340192.5A CN109523528B (en) | 2018-11-12 | 2018-11-12 | Power transmission line extraction method based on unmanned aerial vehicle binocular vision SGC algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109523528A true CN109523528A (en) | 2019-03-26 |
| CN109523528B CN109523528B (en) | 2021-07-13 |
Family
ID=65773390
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811340192.5A Active CN109523528B (en) | 2018-11-12 | 2018-11-12 | Power transmission line extraction method based on unmanned aerial vehicle binocular vision SGC algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109523528B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110288659A (en) * | 2019-05-27 | 2019-09-27 | 魏运 | A kind of Depth Imaging and information acquisition method based on binocular vision |
| CN110473246A (en) * | 2019-07-24 | 2019-11-19 | 哈尔滨工程大学 | A kind of distance measuring method of the multi-overlapped target based on binocular vision |
| CN110738618A (en) * | 2019-10-14 | 2020-01-31 | 河海大学常州校区 | irregular windrow volume measurement method based on binocular camera |
| CN112258604A (en) * | 2020-11-18 | 2021-01-22 | 广州极飞科技有限公司 | Suspended cable reconstruction method and device and electronic equipment |
| CN112770046A (en) * | 2020-12-21 | 2021-05-07 | 深圳市瑞立视多媒体科技有限公司 | Generation method of control SDK of binocular USB camera and control method of binocular USB camera |
| CN113487679A (en) * | 2021-06-29 | 2021-10-08 | 哈尔滨工程大学 | Visual ranging signal processing method for automatic focusing system of laser marking machine |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103500451A (en) * | 2013-10-10 | 2014-01-08 | 中国科学院上海技术物理研究所 | Independent floating ice extraction method for satellite data |
| CN103679707A (en) * | 2013-11-26 | 2014-03-26 | 西安交通大学 | Binocular camera disparity map based road obstacle detection system and method |
| CN107886523A (en) * | 2017-11-01 | 2018-04-06 | 武汉大学 | Vehicle target movement velocity detection method based on unmanned plane multi-source image |
| CN109410207A (en) * | 2018-11-12 | 2019-03-01 | 贵州电网有限责任公司 | A kind of unmanned plane line walking image transmission line faultlocating method based on NCC feature |
-
2018
- 2018-11-12 CN CN201811340192.5A patent/CN109523528B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103500451A (en) * | 2013-10-10 | 2014-01-08 | 中国科学院上海技术物理研究所 | Independent floating ice extraction method for satellite data |
| CN103679707A (en) * | 2013-11-26 | 2014-03-26 | 西安交通大学 | Binocular camera disparity map based road obstacle detection system and method |
| CN107886523A (en) * | 2017-11-01 | 2018-04-06 | 武汉大学 | Vehicle target movement velocity detection method based on unmanned plane multi-source image |
| CN109410207A (en) * | 2018-11-12 | 2019-03-01 | 贵州电网有限责任公司 | A kind of unmanned plane line walking image transmission line faultlocating method based on NCC feature |
Non-Patent Citations (6)
| Title |
|---|
| 张令涛 等: "一种基于图割的改进立体匹配算法", 《机器人》 * |
| 曾纪国 等: "快速鲁棒的立体匹配方法", 《计算机工程与应用》 * |
| 杨浩 等: "基于三维重建的绝缘子覆冰图像监测", 《电力自动化设备》 * |
| 赖小波: "机器人双目立体视觉若干关键理论问题及其技术实现研究", 《中国博士学位论文全文数据库信息科技辑》 * |
| 颜轲 等: "基于图像分割的立体匹配算法", 《计算机应用》 * |
| 颜轲: "数字图像三维立体显示算法研究与实现", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110288659A (en) * | 2019-05-27 | 2019-09-27 | 魏运 | A kind of Depth Imaging and information acquisition method based on binocular vision |
| CN110288659B (en) * | 2019-05-27 | 2021-07-13 | 魏运 | Depth imaging and information acquisition method based on binocular vision |
| CN110473246A (en) * | 2019-07-24 | 2019-11-19 | 哈尔滨工程大学 | A kind of distance measuring method of the multi-overlapped target based on binocular vision |
| CN110473246B (en) * | 2019-07-24 | 2023-10-13 | 哈尔滨工程大学 | Distance measurement method of multiple shielding targets based on binocular vision |
| CN110738618A (en) * | 2019-10-14 | 2020-01-31 | 河海大学常州校区 | irregular windrow volume measurement method based on binocular camera |
| CN110738618B (en) * | 2019-10-14 | 2023-05-16 | 河海大学常州校区 | Irregular stacking volume measuring method based on binocular camera |
| CN112258604A (en) * | 2020-11-18 | 2021-01-22 | 广州极飞科技有限公司 | Suspended cable reconstruction method and device and electronic equipment |
| CN112258604B (en) * | 2020-11-18 | 2023-02-14 | 广州极飞科技股份有限公司 | Suspended cable reconstruction method and device and electronic equipment |
| CN112770046A (en) * | 2020-12-21 | 2021-05-07 | 深圳市瑞立视多媒体科技有限公司 | Generation method of control SDK of binocular USB camera and control method of binocular USB camera |
| CN113487679A (en) * | 2021-06-29 | 2021-10-08 | 哈尔滨工程大学 | Visual ranging signal processing method for automatic focusing system of laser marking machine |
| CN113487679B (en) * | 2021-06-29 | 2023-01-03 | 哈尔滨工程大学 | Visual ranging signal processing method for automatic focusing system of laser marking machine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109523528B (en) | 2021-07-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109523528A (en) | A kind of transmission line of electricity extracting method based on unmanned plane binocular vision SGC algorithm | |
| CN108389256B (en) | Two-three-dimensional interactive unmanned aerial vehicle electric power tower inspection auxiliary method | |
| Sun et al. | Aerial 3D building detection and modeling from airborne LiDAR point clouds | |
| CN106340009B (en) | A kind of electric power line detecting method and system based on parallel binocular | |
| CN103913131B (en) | Free curve method vector measurement method based on binocular vision | |
| CN106709950A (en) | Binocular-vision-based cross-obstacle lead positioning method of line patrol robot | |
| CN110288586A (en) | A kind of multiple dimensioned transmission line of electricity defect inspection method based on visible images data | |
| CN115272271A (en) | Pipeline defect detecting and positioning ranging system based on binocular stereo vision | |
| CN105913013A (en) | Binocular vision face recognition algorithm | |
| CN114782626A (en) | Transformer substation scene mapping and positioning optimization method based on laser and vision fusion | |
| CN103198477A (en) | Apple fruitlet bagging robot visual positioning method | |
| CN107885224A (en) | Unmanned plane barrier-avoiding method based on tri-item stereo vision | |
| CN103729620B (en) | A kind of multi-view pedestrian detection method based on multi-view Bayesian network | |
| CN111126116A (en) | Unmanned ship river channel garbage identification method and system | |
| CN104182968A (en) | Method for segmenting fuzzy moving targets by wide-baseline multi-array optical detection system | |
| CN101765019A (en) | Stereo matching algorithm for motion blur and illumination change image | |
| CN111080685A (en) | Airplane sheet metal part three-dimensional reconstruction method and system based on multi-view stereoscopic vision | |
| CN108399630B (en) | Method for quickly measuring distance of target in region of interest in complex scene | |
| CN116823816B (en) | Detection equipment and detection method based on security monitoring static memory | |
| CN112329723A (en) | Binocular camera-based multi-person human body 3D skeleton key point positioning method | |
| CN104867129A (en) | Light field image segmentation method | |
| Ma et al. | Rapid estimation of apple phenotypic parameters based on 3D reconstruction | |
| CN111680756A (en) | Binocular stereoscopic vision accurate matching method for optimizing inclined plane | |
| CN110009726A (en) | A method of according to the structural relation between plane primitive to data reduction plane | |
| Sun et al. | A vision system based on TOF 3D imaging technology applied to robotic citrus harvesting |
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 |