CN103810462B - High voltage transmission line detection method based on linear targets - Google Patents
High voltage transmission line detection method based on linear targets Download PDFInfo
- Publication number
- CN103810462B CN103810462B CN201210456885.7A CN201210456885A CN103810462B CN 103810462 B CN103810462 B CN 103810462B CN 201210456885 A CN201210456885 A CN 201210456885A CN 103810462 B CN103810462 B CN 103810462B
- Authority
- CN
- China
- Prior art keywords
- line segment
- transmission line
- linear
- hessian matrix
- little
- 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.)
- Expired - Fee Related
Links
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a high voltage transmission line detection method based on linear targets. The high voltage transmission line detection method based on the linear targets comprises the following steps of firstly obtaining a Hessian matrix feature vector direction which is corresponding to every pixel point through a Hessian matrix and enabling the direction to be the direction of the linear target which is formed by the point; extracting points which are approximate to the Hessian matrix feature vector direction through a region growing method to form into small line segments (the linear targets); performing small line segment statistic through a statistical method to obtain a slope K and an intercept B of a power transmission line approximate fitting linear equation and confirming the approximate fitting linear equation of the power transmission line; connecting the small line segments nearby the approximate fitting linear equation and obtaining an accurate position of the high voltage transmission line. The high voltage transmission line detection method based on the linear targets has the advantages of being rapid in detection speed and high in detection accuracy, solving the technical problem of routing inspection of the high voltage transmission line through an unmanned aerial vehicle under complicated background and satisfying the routing inspection requirement of the unmanned aerial vehicle due to the fact that the Hessian matrix is used for detecting of the high voltage transmission line.
Description
Technical field
The present invention relates to image procossing, pattern recognition and Target Recognition, it is exactly specifically using at image
Reason technology for detection and identification high voltage transmission line, belong to linear target identification field.
Background technology
The intact of power circuit is transmission electric energy, the premise of the electricity consumption that ensures safety, so ensureing that the intact of power circuit is
One extremely important task, power circuit polling work is particularly necessary.Traditional power circuit polling flow process is staff parent
From to on-the-spot make an inspection tour circuit.Therefore, patrolling and examining is affected by excessive anthropic factor, in the life of hazardous location entail dangers to line walking workman
Life safety, and manual entry data volume is big, easily malfunction in the manual Input Process of data.Unmanned plane, is a kind of by radio
Remote control equipment or the unmanned vehicle of itself presetting apparatus manipulation.Install in unmanned plane front end and there is stabilizing power
Camera pan-tilt, unmanned plane flies above the side of high-tension line or side first, after starting airborne visual system, using shooting
Machine obtains the video information of scene, and application image technology of identification data integration technology is carried out to circuit and equipment deficiency certainly afterwards
Dynamic detection and analysis.The detection of unmanned plane line walking improves precision and efficiency of detecting, realizes the automatically auxiliary of the defect to line facility
Help checkout and diagnosis and assessment.
At present, linear target recognition methodss substantially can be divided into wire detection method based on edge feature and be based on region
Wire detection method two class.
The present Research of the wire detection based on edge feature：In the linear target recognition methodss based on image border,
In airfield runway is identified, the suitable the Liao Dynasty of poplar waits and is based on mathematical morphology pretreatment, at the beginning of the Freeman Chain-Code-Method with improvement and expansion
Step extracts straight line, then extracts edge line with Hough transform, realizes the automatic identification of airfield runway in image（Poplar is along the Liao Dynasty, Lu
The automatic identification research of airfield runway in the .SAR images such as icepro. Wuhan University of Technology journal .2003,30 (1):56-59）.Luo Jun
Deng by the edge extracting of Sobel operator, refinement and tracking, then carry out fitting a straight line and correction, then extract and meet runway
The straight line pair of condition（Luo Jun, Yang Wei equality. the automatic target detection of airfield runway in infrared image. infrared technique .2003,25
(3):13-17）.Shu Haiyan etc. identification airfield runway be also first pass through Sobel operator carry out edge extracting, refinement after carry out one
The Curvilinear Search of series, connection, carry out after the acquisition of linear search, judgement, merging, connection and parallel lines pair judging identification
（Shu Haiyan. the research of images steganalysis technology and application. Xi'an:Northwestern Polytechnical University .2002）.
Know in method for distinguishing based on the airfield runway of parameter space conversion, Finch etc. passes through to detect Hough transform space
Amplitude in interior same angle, certain distance, to detect the corresponding straight line of airfield runway street lamp sequence, thus identifies airfield runway
（I Finch,A.Antonacopoulos.Identification of airfield runways insynthetic
aperture radar images.Proceedings of the 14th International Conference
onPattern Recognition(ICPR’98).1998,vol.2,1633-1636）.Liu Guangzhi etc. proposes based on Hough
The forward sight airfield runway recognizer of conversion（Liu Guangzhi, Li Jianxun etc. the forward sight airfield runway based on Improved Hough Transform is known
Other method. computer engineering .2004:143-145）.This algorithm utilizes linear edge distinctive gray scale direction and rectilinear direction phase
Consistent feature improves the traditional Hough transform method based on binaryzation marginal point it is achieved that the airport under complex background is run
Road identifies.Bao Fumin proposes a kind of automatic identifying method of airfield runway in diameter radar image（Bao Fumin, Li Aiguo
Deng. the automatic identification of airfield runway in diameter radar image. XI AN JIAOTONG UNIVERSITY Subject Index .2004:1243-1246）.Pass through
Radon conversion is carried out to the image outline extracting in original image, retains maximum several Radon conversion coefficients, then carry out
Radon inverse transformation filters out most of interference lines, only retains the main straight bar in image, is finally filtered using Linear Template
Residual interference lines, and airfield runway is gone out according to threshold test.
The present Research of the wire detection based on provincial characteristicss：Detection method based on edge feature only make use of pixel
Half-tone information, without considering other information in image, therefore has certain limitation.And the target detection based on provincial characteristicss
Method considers not only the half-tone information of pixel, but also include the second-order statisticses of image, cumulative measurement, fractal characteristic,
The features such as grey scale change seriality.
Carry out linear target based on region segmentation and know method for distinguishing being mainly used in bridge, road etc. and embodying in the picture
The target of stronger area information.Domestic many research institutions adopt this method to know method for distinguishing as linear target.?
A kind of method extracting bridge in infrared image is described in the article of Sun Qi（Sun Qi, Cao Zhiguo etc. infrared based on framework
Bridge object identifies. Central China University of Science and Technology journal .Vol.29, No.4, Apr.2001）.The width of bridge is supposed in his method
Spend and only have 2-3 pixel width, and employ a kind of template of seven ranks and bridge area is extracted from background image.From mould
The method that plate extracts bridge area is very ingenious, but this method has the strict restriction of comparison to the width of bridge floor.Be not suitable for pushing away
Wide use.
Based in region method, using the method segmentation of level set and active profile, identify linear target both at home and abroad
A lot of work are done.Keaton T, Ravanbakhsh M et al. proposes to extract the road of high-resolution remote sensing image using level set
Road target, to suppress road topology complex structure, scene information enriches and waits influence of noise（Keaton T,Brokish J.A
level set method for the extraction of roads frommultispectral imagery[C]
.Proceedings of the 31st applied imagery pattern recognitionworkshop,2002:
141-148）.A kind of remote sensing images method for extracting roads based on improvement Level Set Method is described in Tang Xiaofen article（Tang Xiao
Sweet smell, Hou Dibo etc. based on the high-resolution remote sensing image road extraction improving Level Set Method. sensor journal .Vol.23,
No.2,2010）, by introducing each channel information of penalty function item and color space, construction one class be based on many spatial informations and
The level set movements equation that need not reinitialize, has good noise immunity.Li little Mao etc. proposes based on shape holding
Active contour model detects linear target, and this model not only can accurately detect line-like area in image and also have very strong
Anti-noise, resistance to deformation and block performance（Li little Mao, Wang Zhifeng, Tang Yandong. the long vertical bar of active contour model is kept based on shape
Detection [J]. computer engineering, 2008,34）.Rochery, M, Jermyn, I.H etc. are also that shape information is incorporated into one newly
Active contour model in extracting the linear target in remote sensing images such as road, river etc.（Rochery,M.,Jermyn,
I.H.and Zerubia,J.2003.Higher order active contours and their applicationto
the detection of line networks in satellite imagery.In Proc.IEEE Workshop
VLSM, at ICCV, Nice, France）.
The main difficulty causing linear target identification is embodied in：It is complete that the image processing process of lower level seldom can provide target
Whole and accurate profile.In the system using edge detecting technology, due to a reality in effect of noise, with gray level image
Edge is corresponding to be frequently not a continuous line segment, but by many fragments（Little line segment）Composition.These fragments are on the one hand
It is likely due to effect of noise so that the seriality at edge is bad, and create larger gray scale mutation at non-edge；Separately
On the one hand, due to the complexity of concrete scene, the surrounding of target there may be trees, the target such as electric pole, vehicle, target itself
Or the covering of the shade that other targets produce, creates some unhelpful edges.These edges not only increased the meter of detection
Calculation amount, and likely result in the detection of mistake.
Content of the invention
In order to solve problem above, it is an object of the invention to proposing one kind there is accuracy, real-time based on wire
The interrupt method of target, effectively solves what unmanned plane was patrolled and examined to high voltage transmission line under complex background
Technical problem.
The present invention be employed technical scheme comprise that for achieving the above object：A kind of high voltage transmission line inspection based on linear target
Survey method, obtains video or the photographic intelligence of scene, to video or photo by the video camera of carrying on unmanned plane or photographing unit
Information carries out image recognition to detect high voltage transmission line, comprises the following steps：
（1）Calculate each pixel Hessian matrix in image；
（2）Calculate eigenvalue and the characteristic vector of each pixel corresponding Hessian matrix；
（3）Extract consistent adjacent in Hessian matrix characteristic vector direction in all pixels point using region growing methods
Pixel, neighbor pixel constitutes little line segment（Linear target）, length and tiltedly is drawn by the start point/end point coordinate of little line segment
Rate；
（4）Calculate slope identical little line segment length accumulated value in little line segment, will be oblique for the line segment corresponding to maximum accumulated value
Rate is set to the slope K of power transmission line；
（5）Extract slope be K all little line segment, by these little line segments obtain every near linear in the picture cut
Away from B, K and B determines the approximate fits linear equation of every high-voltage line；
（6）Search slope, intercept consistent with K, B little respectively near the approximate fits linear equation of every near linear
Line segment, and these little line segment head and the tail have been connected and composed the accurate location of high voltage transmission line.
The computational methods of the Hessian matrix of described pixel are:With s as yardstick, in image a certain pixel p0=(x0,
y0) place Hessian matrix
Wherein L (x, y) is
Image intensity value.
At the eigenvalue of described Hessian matrix and characteristic vector respectively this pixel of video image, second dervative is big
Little and direction.
The eigenvalue of described Hessian matrix is λ1, λ2, characteristic vector is u1, u2；If | λ1|≤|λ2|, when a certain pixel
When in linear target region, λ1Value is very little（It is ideally 0）；λ2＜ 0, background is dark, and target is bright;λ2＞ 0, then phase
Instead.
The direction of the characteristic vector of described Hessian matrix is the direction of the linear target being made up of pixel, including：If
|λ1|≤|λ2|, then the direction of characteristic vector u1 is the direction of linear target；Conversely, characteristic vector u2 is then linear target
Direction.
Described region growing methods include：
3.1）Assume P point as sub-pixel point, and record the characteristic vector direction of the Hessian matrix of P point；
3.2）Check the characteristic vector direction of the Hessian matrix of 8 neighborhood points of P point, if direction is consistent, labelling should
Dotted state is identical with P dotted state；
3.3）Search for all pixels point successively and record the state of each point, state identical point is sorted out.
Described step（4）Comprise the following steps：
4.1）Obtain the angle of little line segment by the slope of little line segment；
4.2）By little line segment, angularly difference is classified；
4.3）By being added up by length value of little for angle identical line segment；
4.4）Corresponding angle when maximum of little line segment accumulation result is the angle of required power transmission line, and gained slope is defeated
The slope K of electric wire.
Described step（5）Comprise the following steps：
5.1）Calculate the intercept of each little line segment slope little line segment consistent with K；
5.2）The length of little line segment equal for intercept integer value is added up；
5.3）Corresponding intercept when maximum of accumulated value is the intercept of high-voltage line, specially takes and near linear number phase
The same corresponding intercept of several larger accumulated value is intercept B of a plurality of power transmission line to be looked for.
The invention has the advantages that and advantage：
1. the present invention solves the technical problem that unmanned plane is patrolled and examined under complex background to high voltage transmission line.Profit first
Obtain the direction of each pixel corresponding Hessian matrix characteristic vector with Hessian matrix, the direction is by this group
The direction of the linear target becoming；Then the method that region increases is utilized to extract the approximate point in Hessian matrix characteristic vector direction
Form little line segment；Utilize statistical method afterwards, these little line segments are counted, obtains power transmission line approximate fits linear equation
K and B is so that cross the little line segment quantity of this straight line at most, superposition length is the longest, that is, determine the approximate fits straight line side of power transmission line
Journey；Finally connect the little line segment near near linear equation, the as exact position of high voltage transmission line.By theoretical simulation with greatly
The experimental verification of amount, it was demonstrated that detection speed of the present invention is fast, accuracy of detection is high, disclosure satisfy that unmanned plane patrols and examines requirement.
2. compared with prior art, present invention employs linear target for the detection of high voltage transmission line, the method detects
Speed is fast, accuracy of detection is high, solves the technical problem that unmanned plane is patrolled and examined under complex background to high voltage transmission line, can
Meet unmanned plane and patrol and examine requirement.
Brief description
Fig. 1 is method of the present invention flow chart；
The original image example that Fig. 2 shoots for unmanned plane；
Fig. 3（a）Extract result images for linear target direction after inframe Hessian matrix wired in Fig. 2；
Fig. 3（b）Extract result images for linear target direction after inframe Hessian matrix wireless in Fig. 2；
Fig. 4（a）For embodiment of the present invention original image；
Fig. 4（b）For Fig. 4（a）Little line segment after the growth of region（Linear target）Extract result images；
Fig. 5 is to Fig. 4（a）The testing result image of medium-high voltage transmission lines.
Specific embodiment
Below in conjunction with the accompanying drawings and embodiment the present invention is described in further detail.
Flow chart is as shown in figure 1, the present invention adopts Hessian matrix to be used for interrupt method.Hessian square
Battle array（Hessian matrix or Hessian）Also referred to as Hesse matrices, are the Second Order Partials of the real-valued function that independent variable is vector
The square matrix of derivative composition.
Specific implementation step is as follows：
（1）Calculate the gradient vector of each pixel and Hessian matrix in image
If regarding image L (x, y) as be made up of pixel two-dimensional coordinate and its corresponding grey scale value three-dimension curved surface, this three
The coordinate of dimension curved surface can be expressed as：
C=(x, y, L) | L=L (x, y) }
Wherein (x, y) is location of pixels coordinate, and L (x, y) is image intensity value；
Consider its partial structurtes, for certain point p in image0=(x0,y0), the second Taylor series formula of its neighborhood is： In formula, s is yardstick, i.e. template size；And HO, sBe respectively with
S is yardstick, in image certain point p0The gradient vector at place and Hessian matrix,
WhereinWithLxx(p0)、Lxy(p0)、Lyx(p0) and Lyy(p0)（In continuous feelings
Under condition, Lxy(p0)=Lyx(p0)）For image L (p0) second dervative, that is,
Yardstick s unit is pixel herein, can be previously set, and selected numerical value is different, and the little line segment thickness of detection is also different.
（2）Calculate eigenvalue and the characteristic vector of each pixel corresponding Hessian matrix
The eigenvalue of Hessian matrix and characteristic vector are respectively described the size of second dervative and direction at image point,
Characteristic vector is mutually orthogonal, is mutually perpendicular in space；For some fixed size s, the eigenvalue of Hessian matrix is
λ1, λ2, corresponding characteristic vector is respectively u1, u2；If | λ1|≤|λ2|, when a certain pixel is located in linear target region, λ1
Value is very little（It is ideally 0, λ2Value its sign larger determines that this target is bright target or dark target.Therefore for
Linear target must meet | λ1| ≈ 0, | λ1|≤|λ2|；If background is dark, target is bright, then λ2＜ 0;Conversely, then λ2＞ 0.
By comparing λ1And λ2Value judging linear target region, and find out the direction of corresponding characteristic vector；If |
λ1|≤|λ2|, then the direction of characteristic vector u1 is the direction of linear target, otherwise u2 is then the direction of linear target.
Extract eigenvalue and the characteristic vector of Hessian matrix, by comparing λ1And λ2Value judging linear target area
Domain, finds out the direction of corresponding characteristic vector.The original image that Fig. 2 shoots for unmanned plane, Fig. 3 (a) is that in Fig. 2, inframe has high-voltage line
Result, Fig. 3 (b) is the result that in Fig. 2, inframe does not have high-voltage line.In result, the length representative of line segment
The size of eigenvalue, the direction of the direction representative feature vector of line segment.As can be seen that there being the region of high-voltage line from result
Line segment compares concentration, and the direction of characteristic vector is also relatively more approximate.
（3）Extract the approximate point in Hessian matrix characteristic vector direction using the method that region increases and form little line segment（Line
Shape target）
After the direction extracting linear target, following process is done to the point of each linear target：
1. assume P point as sub-pixel point, and record the characteristic vector direction of the Hessian matrix of P point；
2. check the characteristic vector direction of the Hessian matrix of 8 neighborhood points of P point, if direction is approximate, that is, with this feature
Within the range of error of vector direction is ± 1 °, then this dotted state of labelling is identical with P dotted state；
3. search for all pixels point successively and record the state of each point.
The close corresponding region of point in direction can be found out with the method, further from this extracted region to line segment
The information such as point, terminal, direction.Fig. 4（a）For embodiment of the present invention original image, region growth results such as Fig. 4（b）Shown.
（4）Seek the slope K of high voltage transmission line
The statistics approximately equalised little line segment length sum of slope, when length sum is maximum, corresponding slope is and transmits electricity
The slope K of line.The calculating process of the slope K of power transmission line：1）The angle of little line segment can be obtained by the slope of little line segment；2）Will be little
Line segment angularly classified by difference, can be classified with 1 ° for standard in the range of -45 °～+45 °, if angle is little
Number carries out round；3）By being overlapped by length value of little for angle identical line segment, i.e. these length value sums；4）
Corresponding angle when maximum of angle identical little line segment stack result is the angle of required power transmission line, can be calculated power transmission line
Slope K.
（5）Seek intercept B of high voltage transmission line
The length sum of the little line segment of intercept identical, the larger several values of length sum in the little line segment for K for the statistics slope
Corresponding intercept is intercept B of several high voltage transmission lines to be determined, has also determined that power transmission line after obtaining K and B
Approximate fits linear equation.This approximate fits linear equation is linear equation in two unknowns y=Kx+B.The solution of intercept B of power transmission line
Process：1）Calculate the intercept of the little line segment slope little line segment approximate with K, approximate scope is error within ± 1 °；2）To cut
Length away from equal little line segment is overlapped, i.e. these length value sums；3）After taking superposition larger when corresponding intercept
Intercept B for several power transmission lines to be looked for.
（6）Find approximate fits linear equation slope and the intercept little line segment approximate with K and B nearby, and by these little lines
Section head and the tail connect the exact position having obtained power transmission line.The slope little line segment consistent with K refer to slope and K phase ratio error be ±
Little line segment within 1 °, the intercept little line segment consistent with B refers to the little line segment that intercept and B phase ratio error are within 20 pixels.
The testing result of high voltage transmission line is as shown in Figure 5.
Why referred to as approximate fits linear equation refers to the approximate fits of power transmission line, approximate fits, and reason has two：1）
During solving K and B, what result took is all approximation；2）Power transmission line is not inherently absolute straight line.The essence of power transmission line
Really why than approximate fits straight line more accurately position, is because that it is joined end to end by the little line segment near fitting a straight line and forms
, and these little line segments are originally on power transmission line.
Claims (6)
1. a kind of interrupt method based on linear target, is obtained by the video camera carrying or photographing unit on unmanned plane
Obtain video or the photographic intelligence of scene, video or photographic intelligence are carried out image recognition to detect high voltage transmission line, its feature exists
In comprising the following steps：
(1) calculate each pixel Hessian matrix in image；
(2) eigenvalue and the characteristic vector of each pixel corresponding Hessian matrix are calculated；
(3) region growing methods are utilized to extract the consistent neighbor in Hessian matrix characteristic vector direction in all pixels point
Point, neighbor pixel constitutes little line segment, draws length and slope by the starting point of little line segment, terminal point coordinate；
(4) calculate slope identical little line segment length accumulated value in little line segment, the line segment slope corresponding to maximum accumulated value is set
Slope K for power transmission line；
(5) extract all little line segment that slope is K, obtain every near linear intercept B in the picture by these little line segments,
K and B determines the approximate fits linear equation of every high-voltage line；
(6) search slope, intercept little line consistent with K, B respectively near the approximate fits linear equation of every near linear
Section, and these little line segment head and the tail have been connected and composed the accurate location of high voltage transmission line；
Described region growing methods include：
3.1) assume P point as sub-pixel point, and record the characteristic vector direction of the Hessian matrix of P point；
3.2) check the characteristic vector direction of the Hessian matrix of 8 neighborhood points of P point, if direction is consistent, this point-like of labelling
State is identical with P dotted state；
3.3) search for all pixels point successively and record the state of each point, state identical point is sorted out；
Described step (4) comprises the following steps：
4.1) angle of little line segment is obtained by the slope of little line segment；
4.2) by little line segment, angularly difference is classified；
4.3) being added up little for angle identical line segment by length value；
4.4) corresponding angle when maximum of little line segment accumulation result is the angle of required power transmission line, and gained slope is power transmission line
Slope K.
2. a kind of interrupt method based on linear target according to claim 1 it is characterised in that：Described
The computational methods of the Hessian matrix of pixel are:With s as yardstick, in image a certain pixel p0=(x0,y0) place
Hessian matrix
WhereinL (x, y) is image
Gray value.
3. a kind of interrupt method based on linear target according to claim 1 it is characterised in that：Described
The eigenvalue of Hessian matrix and characteristic vector are respectively the size of second dervative and direction at this pixel of video image.
4. a kind of interrupt method based on linear target according to claim 3 it is characterised in that：Described
The eigenvalue of Hessian matrix is λ1, λ2, characteristic vector is u1, u2；If | λ1|≤|λ2|, when a certain pixel is located at linear target
When in region, λ1It is worth very little (ideally for 0)；λ2＜ 0, background is dark, and target is bright；λ2＞ 0, then on the contrary.
5. a kind of interrupt method based on linear target according to claim 1 it is characterised in that：Described
The direction of the characteristic vector of Hessian matrix is the direction of the linear target being made up of pixel, including：If | λ1|≤|λ2|, then
The direction of characteristic vector u1 is the direction of linear target；Conversely, characteristic vector u2 is then the direction of linear target.
6. a kind of interrupt method based on linear target according to claim 1 it is characterised in that：Described
Step (5) comprises the following steps：
5.1) calculate the intercept of each little line segment slope little line segment consistent with K；
5.2) length of little line segment equal for intercept integer value is added up；
5.3) corresponding intercept when maximum of accumulated value is the intercept of high-voltage line, specially takes the same number of near linear
The corresponding intercept of several larger accumulated value is intercept B of a plurality of power transmission line to be looked for.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210456885.7A CN103810462B (en) | 2012-11-14 | 2012-11-14 | High voltage transmission line detection method based on linear targets |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210456885.7A CN103810462B (en) | 2012-11-14 | 2012-11-14 | High voltage transmission line detection method based on linear targets |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103810462A CN103810462A (en) | 2014-05-21 |
| CN103810462B true CN103810462B (en) | 2017-02-08 |
Family
ID=50707210
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210456885.7A Expired - Fee Related CN103810462B (en) | 2012-11-14 | 2012-11-14 | High voltage transmission line detection method based on linear targets |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103810462B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105987665B (en) * | 2015-02-05 | 2020-04-07 | 国家电网公司 | Early warning monitoring device and method for sag point change range of ultra-high voltage transmission line |
| CN104794734B (en) * | 2015-03-30 | 2017-10-31 | 广西大学 | The vision track navigation method of parallel lines cluster in three dimensions |
| CN104751145B (en) * | 2015-04-10 | 2017-10-13 | 西安电子科技大学 | The SAR image electric power line detecting method that local Hough transformation optimizes with morphology |
| CN105450950B (en) * | 2015-12-07 | 2018-07-27 | 北京航空航天大学 | Unmanned plane video jitter removing method |
| CN106127105A (en) * | 2016-06-13 | 2016-11-16 | 长安大学 | Unmanned plane vision patrolling method based on the most random Radon conversion |
| CN106600580B (en) * | 2016-11-30 | 2020-01-21 | 浙江大学 | Hough transform-based power line anomaly identification method and system |
| CN109325935B (en) * | 2018-07-24 | 2022-02-22 | 国网浙江省电力有限公司杭州供电公司 | Power transmission line detection method based on unmanned aerial vehicle image |
| CN112236727A (en) * | 2019-07-30 | 2021-01-15 | 深圳市大疆创新科技有限公司 | Unmanned aerial vehicle simulated flight method and device and recording medium |
| CN110619146B (en) * | 2019-08-12 | 2023-02-14 | 河北工业大学 | Polycrystalline silicon cell crack defect detection method based on structural similarity measurement |
| CN111404075B (en) * | 2020-03-13 | 2021-06-29 | 国家电网有限公司 | Automatic power transmission line measurement method for unmanned aerial vehicle inspection |
| CN112180947B (en) * | 2020-10-22 | 2023-09-12 | 湖南格兰博智能科技有限责任公司 | Method and equipment for selecting initial travelling direction of mobile robot |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101425186A (en) * | 2008-11-17 | 2009-05-06 | 华中科技大学 | Liver subsection method based on CT image and system thereof |
| CN102324109A (en) * | 2011-09-26 | 2012-01-18 | 上海理工大学 | Method for three-dimensionally segmenting insubstantial pulmonary nodule based on fuzzy membership model |
| CN102750703A (en) * | 2012-06-27 | 2012-10-24 | 北京航空航天大学 | Remote sensing image airport automatic detecting method based on linear cutting |
-
2012
- 2012-11-14 CN CN201210456885.7A patent/CN103810462B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101425186A (en) * | 2008-11-17 | 2009-05-06 | 华中科技大学 | Liver subsection method based on CT image and system thereof |
| CN102324109A (en) * | 2011-09-26 | 2012-01-18 | 上海理工大学 | Method for three-dimensionally segmenting insubstantial pulmonary nodule based on fuzzy membership model |
| CN102750703A (en) * | 2012-06-27 | 2012-10-24 | 北京航空航天大学 | Remote sensing image airport automatic detecting method based on linear cutting |
Non-Patent Citations (1)
| Title |
|---|
| 一种面向巡线无人机高压线实时检测与识别算法;赵利坡，范慧杰，朱琳琳，唐延东;《小型微型计算机系统》;20120430;第33卷(第4期);摘要，第882页左栏倒数第2、3行，第883页2.1节，图1 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103810462A (en) | 2014-05-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103810462B (en) | High voltage transmission line detection method based on linear targets | |
| WO2018028103A1 (en) | Unmanned aerial vehicle power line inspection method based on characteristics of human vision | |
| CN105373135B (en) | A kind of method and system of aircraft docking guidance and plane type recognition based on machine vision | |
| CN103839065B (en) | Extraction method for dynamic crowd gathering characteristics | |
| CN110175576A (en) | A kind of driving vehicle visible detection method of combination laser point cloud data | |
| CN103729842B (en) | Based on the fabric defect detection method of partial statistics characteristic and overall significance analysis | |
| CN102073846B (en) | Method for acquiring traffic information based on aerial images | |
| CN101339601B (en) | License plate Chinese character recognition method based on SIFT algorithm | |
| CN105404857A (en) | Infrared-based night intelligent vehicle front pedestrian detection method | |
| CN104951775A (en) | Video technology based secure and smart recognition method for railway crossing protection zone | |
| CN110309701B (en) | Pedestrian re-identification method based on same cross-view-angle area | |
| CN102393961A (en) | Computer vision-based real-time detecting and tracking method for electric power transmission circuit of aircraft | |
| CN104376334B (en) | A kind of pedestrian comparison method of multi-scale feature fusion | |
| CN104281857A (en) | Insulator detecting method based on orthogonal direction projection | |
| Zhang et al. | Automatic diagnosis system of transmission line abnormalities and defects based on UAV | |
| CN103578105B (en) | A kind of multi-source foreign peoples method for registering images based on provincial characteristics | |
| CN103149603B (en) | Road weather detection method based on video | |
| CN110321855A (en) | A kind of greasy weather detection prior-warning device | |
| CN114973116A (en) | Method and system for detecting foreign matters embedded into airport runway at night by self-attention feature | |
| Yuan et al. | Identification method of typical defects in transmission lines based on YOLOv5 object detection algorithm | |
| CN105447431B (en) | A kind of docking aircraft method for tracking and positioning and system based on machine vision | |
| Huang | Research on license plate image segmentation and intelligent character recognition | |
| CN112396572B (en) | Composite insulator double-light fusion method based on feature enhancement and Gaussian pyramid | |
| CN103810494B (en) | Method for high-voltage transmission tower positioning based on linear targets | |
| Wang et al. | A detection method for transmission line insulators based on an improved FCM algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170208 Termination date: 20181114 |