CN107680095A - The electric line foreign matter detection of unmanned plane image based on template matches and optical flow method - Google Patents
The electric line foreign matter detection of unmanned plane image based on template matches and optical flow method Download PDFInfo
- Publication number
- CN107680095A CN107680095A CN201711007961.5A CN201711007961A CN107680095A CN 107680095 A CN107680095 A CN 107680095A CN 201711007961 A CN201711007961 A CN 201711007961A CN 107680095 A CN107680095 A CN 107680095A
- Authority
- CN
- China
- Prior art keywords
- nbsp
- foreign matter
- transmission line
- image
- line
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses the electric line foreign matter detection method based on template matches and the unmanned plane image of optical flow method, belong to picture Processing Technique field.Including transmission line of electricity video image acquisition, the enhancing of line style contrast on border, mean filter, edge line feature extraction, the setting of transmission line of electricity region, template matches, optical flow computation, light stream Binary Sketch of Grey Scale Image, the step such as feature point detection tracking and foreign matter identification.Feature of the invention according to power transmission line level of approximation straight line, it is proposed that line style edge contrast enhancement process and edge line feature extracting method, can accurately extract the edge line of power transmission line, adaptability and strong antijamming capability.According to the big principle of the moving displacement of the object nearer apart from camera corresponding pixel points in the picture, integrated template matching, optical flow computation and image binaryzation propose a kind of foreign matter area determination method, and the degree of accuracy is high, strong robustness.By foreign matter Feature Points Matching track identification foreign matter, the degree of accuracy is high, strong adaptability.
Description
Technical field
Patent of the present invention is related to picture Processing Technique field, more particularly to a kind of based on nobody of template matches and optical flow method
The electric line foreign matter detection method of machine image.
Background technology
In recent years, as the development of power system, transmission line of electricity scope constantly expand, because the foreign matter such as polybag, kite
It is suspended on transmission line malfunction caused by power transmission line frequently to occur, strong influence and damage is caused to people's lives production
Lose.So being monitored transmission line of electricity suspension foreign matter situation, find and fix a breakdown to have great significance in time.
At present, the method for ultra-high-tension power transmission line foreign bodies detection mainly passes through and manually investigated, but because residing for transmission line of electricity
Scope is wide, geographical position is complicated, and artificial difficulty of investigating is larger, and cost is high, and less efficient.Inspection is carried out using unmanned plane, is led to
The transmission line of electricity associated video image for crossing shooting judges transmission line status, has cost low, efficiency high, is not limited by landform
Advantage, it is a kind of effective method of electric line foreign matter monitoring.
The situation for adhering to foreign matter on transmission line of electricity is investigated using the unmanned plane image foreign matter detection method based on intelligent video,
Have the advantages that real-time, accuracy is high and detection speed is fast.It is domestic at present very rare on patent in this respect, in
State's patent " the electric line foreign matter detection method of view-based access control model significance analysis "（Application number：201610269864.2）It is proposed logical
Significance computation model is crossed to judge to define region of interest with the presence or absence of abnormal to analyze power transmission line attachment foreign matter situation
Method, there is the advantages that detection speed is fast, efficiency high, although this method improves the robust of foreign matter judgement to a certain extent
Property, but situations such as background complexity is with foreign matter wide variety is directed to, simple vision significance abnormality detection can not be solved thoroughly
Certainly because caused by the complexity of environment situations such as error in judgement, therefore its accuracy and adaptability all need to be improved.
The content of the invention
For overcome the deficiencies in the prior art, the present invention proposes a kind of unmanned plane figure based on template matches and optical flow method
The electric line foreign matter detection method of picture.A kind of contrast on border Enhancement Method is proposed, effectively strengthens power transmission line edge contrast
Degree, it is convenient to be provided for subsequent edges.A kind of edge detection method is proposed with reference to gray scale morphology principle, can effectively be extracted
Power transmission line edge simultaneously removes background pixel.The video image of unmanned plane inspection shooting is handled, according near apart from camera
Scenery corresponding pixel points in the picture movement velocity it is fast the characteristics of, add range information, with reference to template matches, optical flow method and
The means such as characteristic matching propose a kind of method for recognizing impurities, and accuracy is high, strong adaptability, substantially increases to complex environment background
Antijamming capability.
The technical solution adopted for the present invention to solve the technical problems is：
The electric line foreign matter detection method of the described unmanned plane image based on template matches and optical flow method, including following step
Suddenly：
Step 1：By unmanned plane inspection transmission line of electricity, in the way of being previously set, transmission line of electricity is gathered by airborne camera
Video image；
Step 2：Strengthen the contrast on border of power transmission line according to the linear edge contrast enhancement process of proposition；
Step 3：Mean filter is carried out to the enhanced transmission line of electricity image of contrast on border using line style template；
Step 4：According to the extraction power transmission line of the gray scale morphology Edge Gradient Feature method based on the specific structure element side of proposition
Edge straight line；
Step 5：Power transmission line edge line, image resolution ratio, unmanned plane and transmission line of electricity distance, unmanned plane according to extracting fly
Scanning frequency degree and video camera frame frequency define transmission line of electricity region；
Step 6：The similarity highest matching area of next two field picture at same time interval is found by template matching algorithm；
Step 7：The transmission line of electricity region of interception previous frame image and the template matches region of next two field picture are adjacent as a pair
Image, the optical flow velocity between adjacent two field picture corresponding pixel points is calculated by optical flow method, and according to optical flow velocity numerical value by small
To big order uniform mapping to 0-255 gray scale interval, the gray-scale map of optical flow velocity is made；
Step 8：The binary-state threshold of optical flow velocity gray level image, Ran Houyong are calculated according to the binary-state threshold computational methods of proposition
Target part after binaryzation is surrounded by rectangular box along edge, and rectangular box region is tentatively considered as into foreign matter area
Domain；
Step 9：By four length of sides in rectangular box region respectively to j pixel is extended around, suitably expand foreign matter region, so
The same area in former adjacent image carries out Feature Points Matching and tracking afterwards, and speed is moved according to the pixel of the characteristic point calculated
Degree judges foreign matter.
The transmission line of electricity video image that gathered in the way of being previously set in described step 1 is to make patrol unmanned machine use up
The fixed vertical range away from ground level and between power transmission line of amount, at the uniform velocity travels along the level of moving towards of transmission line of electricity, makes airborne
The axis direction of camera photocentre moves towards perpendicular to transmission line of electricity, timing acquiring transmission line of electricity associated video image.
Linear edge contrast enhancement process in described step 2, specifically includes following steps：
Step A：One layer of wavelet decomposition is carried out to transmission line of electricity image, is divided into approximation component, level detail component, vertical detail point
Amount and diagonal detail component；
Step B：The line style template for being n with length travels through level detail component image and vertical detail component image respectively, judges
Vertical component cv and horizontal component ch in all pixels point of line style template covering at same coordinate position meet condition cv
(i,j)>T1&& ch(i,j)<T2Number of pixels N, wherein threshold value T1According to the gray value that power transmission line edge-perpendicular component is minimum
The amplitude determination of variable gradient, T2Determined according to the amplitude of the maximum gray-value variation gradient of power transmission line edge horizontal component；
Step C：When the number of pixels for the condition that meets is more than N/2, template center's grey scale pixel value of vertical detail component is multiplied
Strengthened with the coefficient more than 1, the coefficient for being multiplied by template center's grey scale pixel value of vertical detail component less than 1 subtracts
It is weak；
Step D：It is right after all pixels point that level detail component image and vertical detail component image have been traveled through according to step BC
Exploded view picture carries out wavelet reconstruction, obtains reconstructed image.
Contrast on border enhancing processing based on wavelet decomposition can increase the gray value ladder of power transmission line edge normal direction
Degree, while make the gray-value variation of edge direction more uniform, it greatly strengthen the accuracy and adaptability of subsequent edges detection.
Line style template in described step 2 and step 3 is rectangle template of the finger widths for a pixel, wherein template
Length is odd number of pixels number, and specific size is according to image resolution ratio sets itself.
It is defeated while blurred background scenery using line style template convolution according to the feature of power transmission line level of approximation straight line
The horizontal direction gray value gradient of electric wire edge changes more uniform and vertical with object edge normal direction gray value ladder
Degree change is little, comparatively, highlights the contrast of object edge, is advantageous to the extraction of power transmission line edge feature.
Specific structure element in described step 4 is using center pixel as starting point, extends vertically upwards p pixel,
It is horizontal to extend the feature templates that q pixel is formed to the right.
The gray scale morphology Edge Gradient Feature method based on specific structure element in described step 4 includes following step
Suddenly：
Step A：With the image after feature structure element traversal mean filter processing, statistics includes hanging down including center pixel
Nogata is to the maximum max1 of p+1 grey scale pixel value, minimum min1 and includes horizontal direction q+1 including center pixel
Maximum max2, the minimum min2 of grey scale pixel value；
Step B：Following computing is carried out according to equation below：
Wherein, grey scale pixel value centered on y, T3And T4Respectively both vertically and horizontally gray-value variation gradient judges threshold
Value, for horizontal direction edge, T3Numerical value is larger, T4It is smaller, can rule of thumb it set；
Step C：Binaryzation is carried out to the image after rim detection, so that it becomes bianry image；
Step D：Straightway in bianry image is extracted by probability Hough line detection algorithm, and it is big according to image resolution ratio
It is small, remove pixel quantity and be less than n1Straightway, and by the absolute value of straight incline angle difference in the range of 0-a degree and line segment
Starting point pixel lateral coordinates difference is in n2Straight-line segments mergence within individual pixel for straight line and extends, and finally gives transmission of electricity
Line edge line.
Edge detection algorithm based on gray scale morphology can be according to the characterizing definition structural element shape of required object edge
State and the requirement for corrode expansion, remove non-targeted edge, the effect of edge extracting is more preferable while object edge is extracted.
Transmission line of electricity region in described step 5 defines method and comprised the following steps：
Step A：N is outwards extended vertically with two straight lines of outermost of the power transmission line edge line detected3After individual pixel
Horizontal linear is horizontal boundary；
Step B：Since picture centre, extend equal length d to both sides in the horizontal directionnVertical line is drawn as power transmission line region
Vertical boundary, wherein, it is ensured that the length 2*d in transmission line of electricity regionnMore than foreign matter pixel within the time of foreign bodies detection
The calculation formula of displacement s, s in the picture are as follows：
Wherein, vertical line direction distances of the L between airborne camera and transmission line of electricity, f are camera focus, v be unmanned plane at the uniform velocity
Flying speed, t be adjacent image frame interval time, n4For the number of image frames of continuous processing required for determination foreign matter.
The similarity of next two field picture by the template matching algorithm searching same intervals time in described step 6 is most
During high matching area, because the offset displacement of consecutive frame respective pixel is smaller, maximum matching template is searched in next two field picture
When, can be on the basis of defined transmission line of electricity region, respectively to extending around n5Individual pixel, then directly in this region
Find the template to match with transmission line of electricity region.
By defining transmission line of electricity region and template matches Search Area, the degree of accuracy of matching can be strengthened, while can also
Enough reduce amount of calculation, increase the calculating speed of algorithm.
Binary-state threshold computational methods in described step 8 comprise the following steps：
Step A：Calculate the optical flow velocity scalar value v of each pixel in gained light stream imagen, then count all pixels point
Average optical flow velocity vaveThe maximum preceding m with optical flow velocitypinThe average optical flow velocity v of individual pixelmave, wherein mpinNumerical value root
According to image resolution ratio and set transmission line of electricity section size sets itself；
Step B：Judge vmaveWith vaveSize, if vmaveV more than 4.5 timesave, then binary-state threshold is set as Te, calculate
Formula is as follows, otherwise judges foreign matter is not present, does not continue to subsequent treatment：
Wherein, vmaxFor the maximum optical flow velocity of all pixels point, as the threshold value T calculatedeWhen not being integer, pass through four houses five
Enter for integer.
In transmission line of electricity image, the translational speed of foreign matter pixel will be far longer than the translational speed of background pixel point,
Because in defined power transmission line region based on background scenery, therefore template matches are mostly the matching of background scenery, for two
Open for matching image, equivalent to background pixel point geo-stationary, foreign matter pixel relative motion.Therefore, optical flow method meter is passed through
The movement velocity of the foreign matter pixel drawn is far longer than the movement velocity of background pixel point, by this principle, according to system
The binary-state threshold computational methods accuracy that meter result defines is high, and being capable of extraordinary segregating foreign objects and background.
The picture element shifting rate for the characteristic point that basis in described step 9 calculates judges that foreign matter comprises the following steps：
Step A：Calculate the displacement ds of consecutive frame foreign matter matching characteristic point according to pixel coordinate in the picture, i.e. consecutive frame
The translational speed v of foreign matter pixel in interval timep1；
Step B：Then again using the template matches region of gained as new transmission line of electricity region, then mould is carried out to next two field picture
The steps such as plate matching, optical flow computation, Feature Points Matching tracking, calculate the translational speed v of foreign matter pixel in consecutive framep2；
Step C：3 trackings are carried out according to step B, calculate the translational speed v of the foreign matter pixel of 3 tracking gainedp1、vp2、
vp3；
Step D：The translational speed v of foreign matter pixel in the picture is calculated according to equation belowypWith background scenery pixel
Translational speed v in imagebp：
Wherein, L1Vertical line direction distance between airborne camera and transmission line of electricity, L2Between airborne camera and ground
Vertical line direction distance, f is camera focus, and v is that unmanned plane flies at a constant speed speed, and t is the interval time of adjacent image frame；
Step E：If judging that 3 characteristic matching regions are all identical, v is calculated respectivelyp1、vp2、vp3With vyp 、vbpError,
If vp1、vp2、vp3Respectively less than vypAnd error is less than 20%, vp1、vp2、vp3It is all higher than vbpAnd error is more than 30%, it is determined that feature
Point is foreign matter characteristic point, i.e., foreign matter is present.
The actual translational speed of foreign matter pixel in the picture is calculated by feature detection and signature tracking, then with basis
The background scenery that formula calculates and the translational speed of foreign matter corresponding pixel points carry out contrast exclusion, can accurately judge very much
The presence of foreign matter, counting accuracy are high.
Compared with prior art, the invention has the advantages that：
1）Contrast on border enhancing processing based on wavelet decomposition can increase the gray value gradient of power transmission line edge normal direction,
Make the gray-value variation of edge direction more uniform simultaneously, greatly strengthen the accuracy and adaptability of subsequent edges detection.
2）According to the feature of power transmission line level of approximation straight line, using line style template convolution, while blurred background scenery,
The horizontal direction gray value gradient of power transmission line edge changes more uniform and vertical with object edge normal direction gray value
Graded is little, comparatively, highlights the contrast of object edge, is advantageous to the extraction of power transmission line edge feature.
3）Edge detection algorithm based on gray scale morphology can be according to the characterizing definition structural element of required object edge
Form and the requirement for corrode expansion, remove non-targeted edge, the effect of edge extracting is more while object edge is extracted
It is good.
4）By defining transmission line of electricity region and template matches Search Area, the degree of accuracy of matching can be strengthened, also simultaneously
Amount of calculation can be reduced, increase the calculating speed of algorithm.
5）According to movement velocity of the camera apart near object corresponding pixel points in the picture it is fast the characteristics of, add away from
From factor, foreign matter region is judged with reference to template matching algorithm, optical flow method and Binarization methods, binaryzation effect is good, can be very
Perfect segregating foreign objects and background.
6）The actual translational speed of foreign matter pixel in the picture is calculated by feature detection and signature tracking, then with root
The background scenery calculated according to formula and the translational speed of foreign matter corresponding pixel points carry out contrast exclusion, can accurately sentence very much
The presence of disconnected foreign matter, counting accuracy are high.
Brief description of the drawings
Fig. 1：The electric line foreign matter detection method flow chart of unmanned plane image based on template matches and optical flow method.
Fig. 2：Line style template based on contrast on border enhancing operation.
Fig. 3：The sectional drawing of transmission line of electricity original gray level image and former gray level image.
Fig. 4：The sectional drawing of transmission line of electricity reconstructed image and reconstructed image.
Fig. 5：Transmission line of electricity image after mean filter.
Fig. 6：Feature structure element template based on gray scale morphology Edge Gradient Feature.
Fig. 7：Transmission line of electricity edge extracting image based on gray scale morphology.
Fig. 8：Transmission line of electricity edge line image.
Fig. 9：The adjacent forms matching image of transmission line of electricity image.
Figure 10：The gray level image of optical flow velocity.
Figure 11：The bianry image of light stream gray level image.
Figure 12：Foreign matter feature point detection and tracking image.
Embodiment
The invention will now be described in further detail with reference to the accompanying drawings：
The electric line foreign matter detection method of unmanned plane image of the invention based on template matches and optical flow method, as shown in figure 1, tool
Body is implemented according to the following steps：
First, the patrol unmanned machine fixed vertical range away from ground level and between power transmission line as far as possible is made, along transmission line of electricity
Move towards level at the uniform velocity to travel, make the axis direction of airborne camera photocentre be moved towards perpendicular to transmission line of electricity, timing acquiring power transmission line
Road associated video image.
Then, contrast on border enhancing processing is carried out to the video image of collection, specifically includes following steps：
Step A：One layer of wavelet decomposition is carried out to transmission line of electricity image, is divided into approximation component, level detail component, vertical detail point
Amount and diagonal detail component；
Step B：Level detail component image and vertical detail component image are traveled through respectively with line style template as shown in Figure 2, are sentenced
Vertical component cv and horizontal component ch in all pixels point of broken string pattern plate covering at same coordinate position meet condition
cv(i,j)>25&& ch(i,j)<4 number of pixels N；
Step C：When the number of pixels for the condition that meets is more than N/2, template center's grey scale pixel value of vertical detail component is multiplied
Strengthened with coefficient 4, being multiplied by coefficient 0.2 to template center's grey scale pixel value of vertical detail component weakens；
Step D：It is right after all pixels point that level detail component image and vertical detail component image have been traveled through according to step BC
Exploded view picture carries out wavelet reconstruction, obtains reconstructed image.
Wherein, the sectional drawing of former gray level image and former gray level image as shown in figure 3, the sectional drawing of reconstructed image and reconstructed image such as
Shown in Fig. 4, contrast two images can be seen that after contrast on border enhancing operation, the ash of power transmission line edge normal direction
Angle value graded increases, and the gray-value variation of edge direction is more uniform.Specific effect increases equivalent at horizontal edge
One chequered with black and white straight line.
Then, average is carried out to the enhanced transmission line of electricity image of contrast on border using line style template as shown in Figure 2
Filtering, as a result as shown in figure 5, as can be seen from Figure, by mean filter, background scenery becomes more to obscure, power transmission line side
Horizontal direction gray value gradient at edge changes more uniform and vertical with object edge normal direction gray value gradient change
Less, comparatively, the contrast of object edge is highlighted, is advantageous to the extraction of power transmission line edge feature.
After mean filter, power transmission line is extracted according to the gray scale morphology Edge Gradient Feature method based on specific structure element
Edge line, specifically include following steps：
Step A：With the image after feature structure element traversal mean filter processing as shown in Figure 6, statistics includes center pixel
The maximum max1, minimum min1 of 3 grey scale pixel values of vertical direction inside and include the level side including center pixel
Maximum max2, minimum min2 to 5 grey scale pixel values；
Step B：Edge power transmission line Edge Gradient Feature is carried out according to equation below：
Wherein, grey scale pixel value centered on y, as a result as shown in Figure 7；
Step C：Binaryzation is carried out to the image after rim detection, so that it becomes bianry image；
Step D：Straightway in bianry image is extracted by probability Hough line detection algorithm, and it is big according to image resolution ratio
It is small, the straightway that pixel quantity is less than 60 is removed, and by the absolute value of straight incline angle difference in the range of 0-5 degree and line segment
Straight-line segments mergence of the starting point pixel lateral coordinates difference within 4 pixels is straight line and extended, and finally gives transmission of electricity
Line edge line, as a result as shown in Figure 8.
As seen from Figure 7, the edge detection algorithm based on gray scale morphology can remove large amount of complex background edge, more
Added with targetedly extracting object edge.After probability Hough straight-line detection, as a result as shown in figure 8, the edge of power transmission line is straight
Line is come out by complete extraction, and non-targeted pixel is completely removed.
After power transmission line edge line extracts, according to power transmission line edge line, image resolution ratio, the unmanned plane extracted
Transmission line of electricity region is defined with transmission line of electricity distance, unmanned plane during flying speed and video camera frame frequency, specifically includes following steps：
Step A：After 150 pixels outwards being extended vertically with two straight lines of outermost of the power transmission line edge line detected
Horizontal linear is horizontal boundary；
Step B：Since picture centre, vertical line is drawn as power transmission line area to both sides extension equal length 200 in the horizontal direction
The vertical boundary in domain, wherein, it is ensured that the length 2*200 in transmission line of electricity region is more than time of the foreign matter pixel in foreign bodies detection
Inside the calculation formula of displacement s, s in the picture is as follows：
Wherein, vertical line direction distances of the L between airborne camera and transmission line of electricity, f are camera focus, v be unmanned plane at the uniform velocity
Flying speed, t be adjacent image frame interval time, n4For the number of image frames of continuous processing required for determination foreign matter.
Then, the similarity highest Matching band of next two field picture at same time interval is found by template matching algorithm
Domain.Wherein,, can be in institute when next two field picture carries out template matches search because consecutive frame pixel moving displacement is smaller
On the basis of defining transmission line of electricity region, respectively to 50 pixels are extended around, then directly find and transmit electricity in this region
The template that land matches.
The transmission line of electricity region of previous frame image and the template matches region of next two field picture are intercepted as a pair of neighbor maps
Picture, as a result as shown in Figure 9.Then the optical flow velocity between adjacent two field picture corresponding pixel points is calculated by optical flow method, and according to
The ascending order uniform mapping of optical flow velocity numerical value makes the gray-scale map of optical flow velocity, as a result to 0-255 gray scale interval
As shown in Figure 10.
Then the threshold value of Binary Sketch of Grey Scale Image is calculated, specifically includes following steps：
Step A：The optical flow velocity scalar value of each pixel in gained light stream image is calculated, then counts the flat of all pixels point
The average optical flow velocity 7.12618 of maximum preceding 100 pixels of equal optical flow velocity 1.02271 and optical flow velocity；
Step B：Judge 7.12618 and 1.02271 size, if 7.12618 are more than the 1.02271 of 4.5 times, set two-value
Change threshold value is Te, calculation formula is as follows, otherwise judges foreign matter is not present, does not continue to subsequent treatment：
Wherein, vmaxFor the maximum optical flow velocity of all pixels point.
Then binaryzation is carried out to light stream gray level image, as a result as shown in figure 11.And with rectangular box along edge by two
Target part after value is surrounded, and rectangular box region is tentatively considered as into foreign matter region.
By four length of sides in rectangular box region respectively to 10 pixels are extended around, suitably expand foreign matter region, so
The same area in former adjacent image carries out Feature Points Matching and tracking afterwards, and speed is moved according to the pixel of the characteristic point calculated
Degree judges foreign matter, specifically includes following steps：
Step A：Calculate the displacement ds of consecutive frame foreign matter matching characteristic point according to pixel coordinate in the picture, i.e. consecutive frame
The translational speed v of foreign matter pixel in interval timep1；
Step B：Then again using the template matches region of gained as new transmission line of electricity region, then mould is carried out to next two field picture
The steps such as plate matching, optical flow computation, Feature Points Matching tracking, calculate the translational speed v of foreign matter pixel in consecutive framep2；
Step C：3 trackings are carried out according to step B, calculate the translational speed v of the foreign matter pixel of 3 tracking gainedp1、vp2、
vp3；
Step D：The translational speed v of foreign matter pixel in the picture is calculated according to equation belowypWith background scenery pixel
Translational speed v in imagebp：
Wherein, L1Vertical line direction distance between airborne camera and transmission line of electricity, L2Between airborne camera and ground
Vertical line direction distance, f is camera focus, and v is that unmanned plane flies at a constant speed speed, and t is the interval time of adjacent image frame；
Step E：If judging that 3 characteristic matching regions are all identical, v is calculated respectivelyp1、vp2、vp3With vyp 、vbpError,
If vp1、vp2、vp3Respectively less than vypAnd error is less than 20%, vp1、vp2、vp3It is all higher than vbpAnd error is more than 30%, it is determined that feature
Point is foreign matter characteristic point, i.e., foreign matter is present.
Hessian threshold values are set as 10000, remarkable characteristic present on foreign matter is detected and chased after with surf algorithms
Track, as a result as shown in figure 12.Mobile speed of the characteristic point in transmission line of electricity image is calculated by the coordinate of pixel in the picture
Degree, the formula then proposed according to above-mentioned steps calculates the translational speed of foreign matter pixel and ground scenery pixel, as a result such as table 1 below
It is shown：
The calculating speed of the actual translational speed of the characteristic point of table 1 and corresponding pixel points
Pass through 3 foreign matter tracing characteristic points, the translational speed v of characteristic point it can be seen from tablep1、vp2、vp3With foreign matter pixel
The calculating speed v of pointypError within 10%, the translational speed v of characteristic pointp1、vp2、vp3It is all higher than the calculating of background pixel point
Speed vbpAnd error is more than vbpOne times, meet above-mentioned regulation, it is possible to determine that, captured by the image on the power transmission line in section
Foreign matter be present.
The preferable embodiment of the present invention is the foregoing is only, is not intended to limit the invention, all spirit in the present invention
Within principle, any modification, equivalent substitution and improvements made etc., it should be included in the scope of the protection.
Claims (10)
1. the electric line foreign matter detection method of the unmanned plane image based on template matches and optical flow method, it is characterised in that described
Method comprises the following steps：
Step 1：By unmanned plane inspection transmission line of electricity, in the way of being previously set, transmission line of electricity is gathered by airborne camera
Video image；
Step 2：Strengthen the contrast on border of power transmission line according to the linear edge contrast enhancement process of proposition；
Step 3：Mean filter is carried out to the enhanced transmission line of electricity image of contrast on border using line style template；
Step 4：According to the extraction power transmission line of the gray scale morphology Edge Gradient Feature method based on the specific structure element side of proposition
Edge straight line；
Step 5：Power transmission line edge line, image resolution ratio, unmanned plane and transmission line of electricity distance, unmanned plane according to extracting fly
Scanning frequency degree and video camera frame frequency define transmission line of electricity region；
Step 6：The similarity highest matching area of next two field picture at same time interval is found by template matching algorithm；
Step 7：The transmission line of electricity region of interception previous frame image and the template matches region of next two field picture are adjacent as a pair
Image, the optical flow velocity between adjacent two field picture corresponding pixel points is calculated by optical flow method, and according to optical flow velocity numerical value by small
To big order uniform mapping to 0-255 gray scale interval, the gray-scale map of optical flow velocity is made；
Step 8：The binary-state threshold of optical flow velocity gray level image, Ran Houyong are calculated according to the binary-state threshold computational methods of proposition
Target part after binaryzation is surrounded by rectangular box along edge, and rectangular box region is tentatively considered as into foreign matter area
Domain；
Step 9：By four length of sides in rectangular box region respectively to j pixel is extended around, suitably expand foreign matter region, so
The same area in former adjacent image carries out Feature Points Matching and tracking afterwards, and speed is moved according to the pixel of the characteristic point calculated
Degree judges foreign matter.
2. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that in described step 1 gathered in the way of being previously set transmission line of electricity video image be make inspection without
The man-machine fixed vertical range away from ground level and between power transmission line as far as possible, is at the uniform velocity travelled along the level of moving towards of transmission line of electricity,
The axis direction of airborne camera photocentre is made to be moved towards perpendicular to transmission line of electricity, timing acquiring transmission line of electricity associated video image.
3. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the linear edge contrast enhancement process in described step 2, specifically include following steps：
Step A：One layer of wavelet decomposition is carried out to transmission line of electricity image, is divided into approximation component, level detail component, vertical detail point
Amount and diagonal detail component；
Step B：The line style template for being n with length travels through level detail component image and vertical detail component image respectively, judges
Vertical component cv and horizontal component ch in all pixels point of line style template covering at same coordinate position meet condition cv
(i,j)>T1&& ch(i,j)<T2Number of pixels N, wherein threshold value T1According to the gray value that power transmission line edge-perpendicular component is minimum
The amplitude determination of variable gradient, T2Determined according to the amplitude of the maximum gray-value variation gradient of power transmission line edge horizontal component；
Step C：When the number of pixels for the condition that meets is more than N/2, template center's grey scale pixel value of vertical detail component is multiplied
Strengthened with the coefficient more than 1, the coefficient for being multiplied by template center's grey scale pixel value of vertical detail component less than 1 subtracts
It is weak；
Step D：It is right after all pixels point that level detail component image and vertical detail component image have been traveled through according to step BC
Exploded view picture carries out wavelet reconstruction, obtains reconstructed image.
4. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the line style template in described step 2 and step 3 is rectangle template of the finger widths for a pixel, its
Middle template length is odd number of pixels number, and specific size is according to image resolution ratio sets itself.
5. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the specific structure element in described step 4 is using center pixel as starting point, extends vertically upwards p picture
Vegetarian refreshments, it is horizontal to extend the feature templates that q pixel is formed to the right.
6. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the gray scale morphology Edge Gradient Feature method based on specific structure element in described step 4 includes
Following steps：
Step A：With the image after feature structure element traversal mean filter processing, statistics includes hanging down including center pixel
Nogata is to the maximum max1 of p+1 grey scale pixel value, minimum min1 and includes horizontal direction q+1 including center pixel
Maximum max2, the minimum min2 of grey scale pixel value；
Step B：Following computing is carried out according to equation below：
<math display = 'block'>
<mrow>
<mi>y</mi>
<mo>=</mo>
<mfenced open = '{' close = ''>
<mtable rowalign='center'>
<mtr>
<mtd>
<mrow>
<mi>min1</mi>
<mtext>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;&nbsp;</mtext>
<mi>abs</mi>
<mo stretchy='false'>(</mo>
<mi>max1</mi>
<mo>&minus;</mo>
<mi>min1</mi>
<mo stretchy='false'>)</mo>
<mo>></mo>
<msub>
<mi>T</mi>
<mn>3</mn>
</msub>
<mo>&</mo>
<mo>&</mo>
<mi>abs</mi>
<mo stretchy='false'>(</mo>
<mi>max2</mi>
<mo>&minus;</mo>
<mi>min2</mi>
<mo stretchy='false'>)</mo>
<mo><</mo>
<msub>
<mi>T</mi>
<mn>4</mn>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>y</mi>
<mtext>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;</mtext>
<mi>otherwise</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
</math>
Wherein, grey scale pixel value centered on y, T3And T4Respectively both vertically and horizontally gray-value variation gradient judges threshold
Value, for horizontal direction edge, T3Numerical value is larger, T4It is smaller, can rule of thumb it set；
Step C：Binaryzation is carried out to the image after rim detection, so that it becomes bianry image；
Step D：Straightway in bianry image is extracted by probability Hough line detection algorithm, and it is big according to image resolution ratio
It is small, remove pixel quantity and be less than n1Straightway, and by the absolute value of straight incline angle difference in the range of 0-a degree and line segment
Starting point pixel lateral coordinates difference is in n2Straight-line segments mergence within individual pixel for straight line and extends, and finally gives transmission of electricity
Line edge line.
7. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the transmission line of electricity region in described step 5 defines method and comprised the following steps：
Step A：N is outwards extended vertically with two straight lines of outermost of the power transmission line edge line detected3Water after individual pixel
Flat line is horizontal boundary；
Step B：Since picture centre, extend equal length d to both sides in the horizontal directionnVertical line is drawn as power transmission line region
Vertical boundary, wherein, it is ensured that the length 2*d in transmission line of electricity regionnMore than foreign matter pixel within the time of foreign bodies detection
The calculation formula of displacement s, s in the picture are as follows：
<math display = 'block'>
<mrow>
<mi>s</mi>
<mo>=</mo>
<mfrac>
<mi>fvt</mi>
<mi>L</mi>
</mfrac>
<mo>&times;</mo>
<msub>
<mi>n</mi>
<mn>4</mn>
</msub>
</mrow>
</math>
Wherein, vertical line direction distances of the L between airborne camera and transmission line of electricity, f are camera focus, v be unmanned plane at the uniform velocity
Flying speed, t be adjacent image frame interval time, n4For the number of image frames of continuous processing required for determination foreign matter.
8. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the phase of next two field picture that the same intervals time is found by template matching algorithm in described step 6
During like degree highest matching area, because the offset displacement of consecutive frame respective pixel is smaller, maximum is searched in next two field picture
, can be on the basis of defined transmission line of electricity region, respectively to extending around n during with template5Individual pixel, then directly exists
Find the template to match with transmission line of electricity region in this region.
9. the electric line foreign matter detection side of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the binary-state threshold computational methods in described step 8 comprise the following steps：
Step A：Calculate the optical flow velocity scalar value v of each pixel in gained light stream imagen, then count all pixels point
Average optical flow velocity vaveThe maximum preceding m with optical flow velocitypinThe average optical flow velocity v of individual pixelmvae, wherein mpinNumerical value root
According to image resolution ratio and set transmission line of electricity section size sets itself；
Step B：Judge vmvaeWith vaveSize, if vmvaeV more than 4.5 timesave, then binary-state threshold is set as Te, meter
It is as follows to calculate formula, otherwise judges foreign matter is not present, does not continue to subsequent treatment：
<math display = 'block'>
<mrow>
<msub>
<mi>T</mi>
<mi>e</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mn>3.5</mn>
<mo>&times;</mo>
<msub>
<mi>v</mi>
<mi>ave</mi>
</msub>
<mo>&times;</mo>
<mn>255</mn>
</mrow>
<msub>
<mi>v</mi>
<mi>max</mi>
</msub>
</mfrac>
</mrow>
</math>
Wherein, vmaxFor the maximum optical flow velocity of all pixels point, as the threshold value T calculatedeWhen not being integer, by rounding up
For integer.
10. the electric line foreign matter detection of the unmanned plane image according to claim 1 based on template matches and optical flow method
Method, it is characterised in that the picture element shifting rate for the characteristic point that the basis in described step 9 calculates judge foreign matter include with
Lower step：
Step A：Calculate the displacement ds of consecutive frame foreign matter matching characteristic point according to pixel coordinate in the picture, i.e. consecutive frame
The translational speed v of foreign matter pixel in interval timep1；
Step B：Then again using the template matches region of gained as new transmission line of electricity region, then mould is carried out to next two field picture
The steps such as plate matching, optical flow computation, Feature Points Matching tracking, calculate the translational speed v of foreign matter pixel in consecutive framep2；
Step C：3 trackings are carried out according to step B, calculate the translational speed v of the foreign matter pixel of 3 tracking gainedp1、vp2、
vp3；
Step D：The translational speed v of foreign matter pixel in the picture is calculated according to equation belowypWith background scenery pixel
Translational speed v in imagebp：
<math display = 'block'>
<mrow>
<msub>
<mi>v</mi>
<mi>yp</mi>
</msub>
<mo>=</mo>
<mfrac>
<mi>fvt</mi>
<msub>
<mi>L</mi>
<mn>1</mn>
</msub>
</mfrac>
</mrow>
</math><math display = 'block'>
<mrow>
<msub>
<mi>v</mi>
<mi>bp</mi>
</msub>
<mo>=</mo>
<mfrac>
<mi>fvt</mi>
<msub>
<mi>L</mi>
<mn>2</mn>
</msub>
</mfrac>
</mrow>
</math>；
Wherein, L1Vertical line direction distance between airborne camera and transmission line of electricity, L2Between airborne camera and ground
Vertical line direction distance, f are camera focus, and v is that unmanned plane flies at a constant speed speed, and t is the interval time of adjacent image frame；
Step E：If judging that 3 characteristic matching regions are all identical, v is calculated respectivelyp1、vp2、vp3With vyp、vbpError, if
vp1、vp2 、vp3Respectively less than vypAnd error is less than 20%, vp1、vp2 、vp3It is all higher than vbpAnd error is more than 30%, it is determined that characteristic point
For foreign matter characteristic point, i.e. foreign matter is present.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711007961.5A CN107680095A (en) | 2017-10-25 | 2017-10-25 | The electric line foreign matter detection of unmanned plane image based on template matches and optical flow method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711007961.5A CN107680095A (en) | 2017-10-25 | 2017-10-25 | The electric line foreign matter detection of unmanned plane image based on template matches and optical flow method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107680095A true CN107680095A (en) | 2018-02-09 |
Family
ID=61142207
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711007961.5A Pending CN107680095A (en) | 2017-10-25 | 2017-10-25 | The electric line foreign matter detection of unmanned plane image based on template matches and optical flow method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107680095A (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108767744A (en) * | 2018-08-23 | 2018-11-06 | 国网江苏省电力有限公司电力科学研究院 | One kind is from the long-range remove device of motion tracking power grid foreign matter laser and sweep-out method |
| CN110223310A (en) * | 2019-05-22 | 2019-09-10 | 上海大学 | A kind of line-structured light center line and cabinet edge detection method based on deep learning |
| CN110349172A (en) * | 2019-06-28 | 2019-10-18 | 华南理工大学 | Transmission line of electricity external force damage prevention method for early warning based on image procossing and binocular solid ranging |
| CN110414308A (en) * | 2019-05-16 | 2019-11-05 | 南京理工大学 | A kind of target identification method for dynamic foreign matter on transmission line of electricity |
| CN110751618A (en) * | 2019-06-05 | 2020-02-04 | 浙江大华技术股份有限公司 | Floater detection method and device and electronic equipment |
| CN110852320A (en) * | 2019-11-08 | 2020-02-28 | 积成电子股份有限公司 | Transmission channel foreign matter intrusion detection method based on deep learning |
| CN110956614A (en) * | 2019-11-11 | 2020-04-03 | 国网山东省电力公司电力科学研究院 | Ground wire foreign matter detection method and device based on iterative search and projection method |
| CN111275697A (en) * | 2020-02-10 | 2020-06-12 | 西安交通大学 | Battery silk-screen quality detection method based on ORB feature matching and LK optical flow method |
| CN111277729A (en) * | 2020-02-26 | 2020-06-12 | 新疆大学 | Video image processing method and device and electronic equipment |
| CN111311567A (en) * | 2020-02-12 | 2020-06-19 | 北京交通大学 | Method for identifying fastener and steel rail diseases of track line image |
| CN111814547A (en) * | 2020-06-03 | 2020-10-23 | 国网浙江嘉善县供电有限公司 | Overhead high-voltage wire tracking detection system and method based on visual identification |
| CN111967335A (en) * | 2020-07-21 | 2020-11-20 | 广东工业大学 | Method for identifying foreign matters on power transmission line based on image processing |
| CN112329584A (en) * | 2020-10-29 | 2021-02-05 | 深圳技术大学 | Method, system and equipment for automatically identifying foreign matters in power grid based on machine vision |
| CN112598623A (en) * | 2020-12-02 | 2021-04-02 | 国网冀北电力有限公司经济技术研究院 | Method and system for collecting line stock by utilizing edge calculation |
| CN112801141A (en) * | 2021-01-08 | 2021-05-14 | 吉林大学 | Heterogeneous image matching method based on template matching and twin neural network optimization |
| CN113673514A (en) * | 2021-08-11 | 2021-11-19 | 国网山东省电力公司微山县供电公司 | Method and system for detecting invasion of foreign matters into power transmission line |
| CN115690430A (en) * | 2022-12-29 | 2023-02-03 | 尚特杰电力科技有限公司 | Photovoltaic module edge feature map supplementing method, system, equipment and storage medium |
| US11630440B2 (en) | 2018-11-08 | 2023-04-18 | Fanuc Corporation | Wire disconnection prediction device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105957077A (en) * | 2015-04-29 | 2016-09-21 | 国网河南省电力公司电力科学研究院 | Detection method for foreign body in transmission lines based on visual saliency analysis |
| CN106649598A (en) * | 2016-11-22 | 2017-05-10 | 云南电网有限责任公司电力科学研究院 | Detection method for icing thickness of power transmission line |
| CN106705866A (en) * | 2016-12-14 | 2017-05-24 | 云南电网有限责任公司电力科学研究院 | Visible light image-based transmission line icing detection method |
| CN106960438A (en) * | 2017-03-25 | 2017-07-18 | 安徽继远软件有限公司 | Method for recognizing impurities to transmission line of electricity is converted based on Hough straight line |
-
2017
- 2017-10-25 CN CN201711007961.5A patent/CN107680095A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105957077A (en) * | 2015-04-29 | 2016-09-21 | 国网河南省电力公司电力科学研究院 | Detection method for foreign body in transmission lines based on visual saliency analysis |
| CN106649598A (en) * | 2016-11-22 | 2017-05-10 | 云南电网有限责任公司电力科学研究院 | Detection method for icing thickness of power transmission line |
| CN106705866A (en) * | 2016-12-14 | 2017-05-24 | 云南电网有限责任公司电力科学研究院 | Visible light image-based transmission line icing detection method |
| CN106960438A (en) * | 2017-03-25 | 2017-07-18 | 安徽继远软件有限公司 | Method for recognizing impurities to transmission line of electricity is converted based on Hough straight line |
Non-Patent Citations (2)
| Title |
|---|
| 金立军 等: ""基于航拍图像的输电线路异物识别"", 《同济大学学报》 * |
| 魏国剑: ""融合光流检测与模板匹配的目标跟踪算法"", 《计算机应用研究》 * |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108767744A (en) * | 2018-08-23 | 2018-11-06 | 国网江苏省电力有限公司电力科学研究院 | One kind is from the long-range remove device of motion tracking power grid foreign matter laser and sweep-out method |
| CN108767744B (en) * | 2018-08-23 | 2024-05-03 | 国网江苏省电力有限公司电力科学研究院 | Laser remote cleaning device and method for automatically tracking foreign matters in power grid |
| US11630440B2 (en) | 2018-11-08 | 2023-04-18 | Fanuc Corporation | Wire disconnection prediction device |
| CN110414308A (en) * | 2019-05-16 | 2019-11-05 | 南京理工大学 | A kind of target identification method for dynamic foreign matter on transmission line of electricity |
| CN110223310A (en) * | 2019-05-22 | 2019-09-10 | 上海大学 | A kind of line-structured light center line and cabinet edge detection method based on deep learning |
| CN110223310B (en) * | 2019-05-22 | 2023-07-18 | 上海大学 | Line structure light center line and box edge detection method based on deep learning |
| CN110751618A (en) * | 2019-06-05 | 2020-02-04 | 浙江大华技术股份有限公司 | Floater detection method and device and electronic equipment |
| CN110349172B (en) * | 2019-06-28 | 2022-12-16 | 华南理工大学 | Power transmission line external damage prevention early warning method based on image processing and binocular stereo ranging |
| CN110349172A (en) * | 2019-06-28 | 2019-10-18 | 华南理工大学 | Transmission line of electricity external force damage prevention method for early warning based on image procossing and binocular solid ranging |
| CN110852320A (en) * | 2019-11-08 | 2020-02-28 | 积成电子股份有限公司 | Transmission channel foreign matter intrusion detection method based on deep learning |
| CN110852320B (en) * | 2019-11-08 | 2022-06-24 | 积成电子股份有限公司 | Transmission channel foreign matter intrusion detection method based on deep learning |
| CN110956614A (en) * | 2019-11-11 | 2020-04-03 | 国网山东省电力公司电力科学研究院 | Ground wire foreign matter detection method and device based on iterative search and projection method |
| CN110956614B (en) * | 2019-11-11 | 2023-04-07 | 国网山东省电力公司电力科学研究院 | Ground wire foreign matter detection method and device based on iterative search and projection method |
| CN111275697A (en) * | 2020-02-10 | 2020-06-12 | 西安交通大学 | Battery silk-screen quality detection method based on ORB feature matching and LK optical flow method |
| CN111275697B (en) * | 2020-02-10 | 2022-04-22 | 西安交通大学 | Battery silk-screen quality detection method based on ORB feature matching and LK optical flow method |
| CN111311567A (en) * | 2020-02-12 | 2020-06-19 | 北京交通大学 | Method for identifying fastener and steel rail diseases of track line image |
| CN111277729A (en) * | 2020-02-26 | 2020-06-12 | 新疆大学 | Video image processing method and device and electronic equipment |
| CN111814547A (en) * | 2020-06-03 | 2020-10-23 | 国网浙江嘉善县供电有限公司 | Overhead high-voltage wire tracking detection system and method based on visual identification |
| CN111967335A (en) * | 2020-07-21 | 2020-11-20 | 广东工业大学 | Method for identifying foreign matters on power transmission line based on image processing |
| CN112329584A (en) * | 2020-10-29 | 2021-02-05 | 深圳技术大学 | Method, system and equipment for automatically identifying foreign matters in power grid based on machine vision |
| CN112598623A (en) * | 2020-12-02 | 2021-04-02 | 国网冀北电力有限公司经济技术研究院 | Method and system for collecting line stock by utilizing edge calculation |
| CN112801141B (en) * | 2021-01-08 | 2022-12-06 | 吉林大学 | Heterogeneous image matching method based on template matching and twin neural network optimization |
| CN112801141A (en) * | 2021-01-08 | 2021-05-14 | 吉林大学 | Heterogeneous image matching method based on template matching and twin neural network optimization |
| CN113673514A (en) * | 2021-08-11 | 2021-11-19 | 国网山东省电力公司微山县供电公司 | Method and system for detecting invasion of foreign matters into power transmission line |
| CN113673514B (en) * | 2021-08-11 | 2024-06-04 | 国网山东省电力公司微山县供电公司 | Foreign matter intrusion detection method and system for power transmission line |
| CN115690430A (en) * | 2022-12-29 | 2023-02-03 | 尚特杰电力科技有限公司 | Photovoltaic module edge feature map supplementing method, system, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107680095A (en) | The electric line foreign matter detection of unmanned plane image based on template matches and optical flow method | |
| CN107045629B (en) | Multi-lane line detection method | |
| CN107679520B (en) | Lane line visual detection method suitable for complex conditions | |
| CN103077384B (en) | A kind of method and system of vehicle-logo location identification | |
| CN104751491B (en) | A kind of crowd's tracking and people flow rate statistical method and device | |
| CN105005771B (en) | A kind of detection method of the lane line solid line based on light stream locus of points statistics | |
| CN103164858B (en) | Adhesion crowd based on super-pixel and graph model is split and tracking | |
| Zhang et al. | Edge detection algorithm of image fusion based on improved Sobel operator | |
| WO2017041396A1 (en) | Driving lane data processing method, device, storage medium and apparatus | |
| CN103546726B (en) | Method for automatically discovering illegal land use | |
| Yuan et al. | Robust lane detection for complicated road environment based on normal map | |
| CN102750708B (en) | Affine motion target tracing algorithm based on fast robust feature matching | |
| CN110379168B (en) | Traffic vehicle information acquisition method based on Mask R-CNN | |
| CN109255350B (en) | New energy license plate detection method based on video monitoring | |
| CN106845364B (en) | Rapid automatic target detection method | |
| CN106530281B (en) | Unmanned plane image fuzzy Judgment method and system based on edge feature | |
| CN106600625A (en) | Image processing method and device for detecting small-sized living thing | |
| CN112801022A (en) | Method for rapidly detecting and updating road boundary of unmanned mine card operation area | |
| CN104835147A (en) | Method for detecting crowded people flow in real time based on three-dimensional depth map data | |
| CN103218816A (en) | Crowd density estimation method and pedestrian volume statistical method based on video analysis | |
| CN107341480A (en) | A kind of crowd massing detection method of modified PCCNN neural network models | |
| CN105404857A (en) | Infrared-based night intelligent vehicle front pedestrian detection method | |
| CN103426164B (en) | Based on scallop size computing method and the scallop sorting system of Opencv graphical analysis | |
| CN102592128B (en) | Method and device for detecting and processing dynamic image and display terminal | |
| CN103971524B (en) | A kind of traffic flow detection method based on machine vision |
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 | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20210806 |
|
| AD01 | Patent right deemed abandoned |