CN103761722B - Fixed wing unmanned aerial vehicle touring image accurately-splicing method for power transmission line - Google Patents
Fixed wing unmanned aerial vehicle touring image accurately-splicing method for power transmission line Download PDFInfo
- Publication number
- CN103761722B CN103761722B CN201410020480.8A CN201410020480A CN103761722B CN 103761722 B CN103761722 B CN 103761722B CN 201410020480 A CN201410020480 A CN 201410020480A CN 103761722 B CN103761722 B CN 103761722B
- Authority
- CN
- China
- Prior art keywords
- transmission line
- power transmission
- image
- wing unmanned
- point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a fixed wing unmanned aerial vehicle touring image accurately-splicing method for a power transmission line. The accurately-splicing method includes the following steps: (1) conducting data preparation and sample training; (2) conducting feature extraction on the power transmission line based on textural features and straight line features to obtain a power transmission line distributing area and power transmission line extraction data, screening the power transmission line extraction data through the power transmission line distributing area, and obtaining power transmission line extracting results through power transmission line double-edge features; (3) conducting SIFT image registration and fusing based on power transmission line customization on the power transmission line extracting results. According to the fixed wing unmanned aerial vehicle touring image accurately-splicing method, curvelet transformation is adopted for achieving extraction on power transmission line texture information and detection on power transmission line distribution, and the power transmission line double-edge features in high-resolution unmanned aerial vehicle images are further adopted for obtaining the reliable and accurate power transmission line extracting results.
Description
Technical field
The present invention relates to image split-joint method, be specifically related to a kind of method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image.
Background technology
Existing fixed-wing unmanned plane is maked an inspection tour image mosaic technology and is mainly used in mapping aspect, and the joining method of employing is numerous, exist subject matter comprise that splicing precision is not high and splicing speed is slow.
Fixed-wing unmanned plane is maked an inspection tour image mosaic aspect and is not applied at present, domesticly also to report without related ends, make an inspection tour sequence image to transmission line of electricity fixed-wing unmanned plane to splice, form complete power transmission line corridor panoramic picture, transmission line of electricity particularly corridor information accurately directly perceived can be provided, fixed-wing unmanned plane during flying height is high, flight generally reaches 300-500m to ground level, and extra high voltage network wire diameter is generally about 30mm, realize the high-precision joining that transmission line of electricity fixed-wing unmanned plane makes an inspection tour image, need to propose a kind of brand-new joining method.
It is higher to splicing accuracy requirement that fixed-wing unmanned plane makes an inspection tour image, wire splicing is a gordian technique point wherein, due to the restriction of prior art, use existing mapping splicing to carry out transmission line of electricity fixed-wing unmanned plane and make an inspection tour image mosaic, there will be obvious problem of misalignment, occurred that wire does not dock, wire moves towards factitious phenomenon.
Summary of the invention
For above deficiency, the object of this invention is to provide a kind of method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image.
For realizing above object, the technical scheme that this invention takes is:
For the method that transmission line of electricity fixed-wing unmanned plane tour image accurately splices, it comprises the following steps:
Step 1, data encasement and sample training; Described step 1 comprises the following steps:
Step 1.1, be the training subimage of multiple formed objects by sample cutting;
Step 1.2, the warp wavelet based on USFFT algorithm is adopted to carry out sample characteristics extraction, to obtain training sample to described training subimage;
Step 1.3, calculated the PCA parameter of described training sample by PCA algorithm, described PCA parameter comprises mean value vector and projection matrix;
The training of step 1.4, RBF neural;
Step 2, based on textural characteristics and linear feature to power transmission line feature extraction, to obtain power transmission line distributed areas and power transmission line extracts data; Extract data by described power transmission line distributed areas to described power transmission line to screen, obtain power transmission line by the dual edge feature of power transmission line and extract result;
Step 3, described power transmission line extracted to SIFT image registration and fusion that result customizes based on power transmission line.
Described step 1.4 comprises the following steps:
To described training subimage in include power transmission line be labeled as 1, what do not comprise power transmission line is labeled as-1;
Setting RBF network parameter, nodes and the number of plies;
With the form of random weight for starting point at least repeats ten neural metwork trainings;
Export the network weight of the RBF neural after having trained.
Described step 2 comprises the following steps:
Step 2.1, textural characteristics is utilized to carry out the detection of power transmission line distributed areas;
Step 2.2, being carried out the detection of image border by Canny operator, is original candidates edge;
Step 2.3, by Probabilistic Hough Transform, straight-line detection is carried out to power transmission line, complete the extraction of power transmission line, and the power transmission line of fracture is connected, preserve linear position information;
Step 2.4, according to power transmission line distributed areas detect carry out of result to Hough straight-line detection screen, the straight line in power transmission line distributed areas is chosen as power transmission line original candidates edge;
Step 2.5, carry out the further screening at power transmission line original candidates edge according to following rule:
A) direction at original candidates edge meets the direction of the power transmission line calculated according to geography information;
B) length at original candidates edge exceedes the threshold value of specifying;
Step 2.6, by the candidate edge of further screening, choose and be parallel to each other and distance two straight lines that are M pixel, getting its average is reference line, described M=power transmission line diameter/image resolution ratio;
If the reference line do not satisfied condition, M=M-1 carries out standard straight-line retrieval again, until M=0 and single straight line;
The result that step 2.7, output power transmission line extract.
Described step 2.1 utilizes textural characteristics to carry out the detection of power transmission line distributed areas and comprises the following steps:
Step 2.1.1, Aerial Images is converted into gray level image;
Step 2.1.2, utilize the subregion of the schema extraction gray level image of moving window;
Step 2.1.3, power transmission line regional determination is carried out to every sub regions;
Step 2.1.4, travel through all subregions, complete the region detection of power transmission line distribution, wherein, area detection result intermediate value be greater than 0 be power transmission line distributed areas.
Described step 2.1.3 comprises the following steps:
Step 2.1.3.1, USFFT algorithm is utilized to extract bent wave characteristic to subregion;
USFFT algorithm key step comprises:
A) Fourier transform that 2DFFT obtains f is done to f
B) under often pair of yardstick and angle (j, l) to f [n
1, n
2] sample (or interpolation) obtain f [n again
1, n
2-n
1tan θ
l];
C) by f [n
1, n
2-n
1tan θ
l] and window function U
j[n
1, n
2] being multiplied obtains f
j,l;
D) by f
j,ldo inverse 2DFFT conversion and obtain bent wave system number c
d(i, l, k).
Step 2.1.3.2, antithetical phrase extracted region to feature carry out based on principal component analysis Feature Dimension Reduction;
Step 2.1.3.3, the RBF neural utilizing training to obtain differentiate the feature after dimensionality reduction, and comprising power transmission line if judge, be 1, otherwise assignment is-1 by subregion assignment;
Described step 3 comprises the following steps:
Step 3.1, the SIFT feature point detection customized based on power transmission line;
Step 3.2, unique point describe vector and generate, and utilize SIFT operator extraction SIFT feature vector, then extract bent wave characteristic describe vector as unique point together with SIFT feature vector;
The coupling of step 3.3, unique point, the coupling of described unique point describes the similarity criteria of the Euclidean distance between vector as Feature Points Matching using unique point, adopt preferential k-d to set and carry out first search to search two approximate KNN unique points of each unique point, calculate the Euclidean distance of two corresponding with it respectively approximate KNN unique points of this unique point, and calculate the ratio of two Euclidean distances, if ratio is less than threshold value, then the match is successful, otherwise, again mate;
Step 3.4, image mosaic and fusion, realize the splicing between two width or multiple image by image weighting average method.
Described step 3.1 comprises the following steps:
On the power transmission line extracted, N number of unique point is selected with SIFT feature point detection algorithm, the point on all power transmission lines is traveled through in graphical rule space, judge the relation of point in itself and neighborhood, if the value of this point be greater than or less than neighborhood value a little, then this point is candidate feature point;
Other regions utilize SIFT feature point detection algorithm to choose N number of unique point equally in the picture, require distant with described candidate feature point and are evenly distributed in image-region as far as possible.
Also Image semantic classification is comprised between described step 1 and step 2.
Described Image semantic classification comprises the following steps:
Utilize the geometric calibration that unmanned plane POS information is carried out Aerial Images, to correct because fixed-wing unmanned plane makes an inspection tour distortion, the distortion of the image that shooting angle and lens distortion cause;
Realize carrying out denoising when fully retaining boundary information to image by anisotropic filtering.
The power transmission line that the present invention mainly proposes based on warp wavelet and dual edge feature extracts the splicing that the SIFT feature matching algorithm customized with wire realizes image, avoid when power transmission line fixed-wing unmanned plane makes an inspection tour Sequential images mosaic and occur comparatively big error problem, decrease mistake and a phenomenon unmatched, complete the quick high accuracy splicing that transmission line of electricity fixed-wing unmanned plane makes an inspection tour sequence image.It utilizes the extraction having anisotropic warp wavelet to realize to power transmission line textural characteristics, the problem that power transmission line extracts is converted into the power transmission line test problems comprising subregion by the thinking utilizing moving window and RBF neural to differentiate, the Integrated Selection that power transmission line extracts result is realized by the dual edge characteristic sum texture information of power transmission line, finally utilize the SIFT operator of wire weight to determine the unique point of image, and fully utilize continuity and integrality that coupling that SIFT feature and bent wave characteristic carry out realization character point ensures wire in image registration.
The present invention compared with prior art, its beneficial effect is: the present invention utilizes warp wavelet to realize the detection distributed to extraction and the power transmission line of power transmission line texture information, also utilize the power transmission line dual edge feature in high resolving power unmanned plane image, obtain and more reliable extract result with power transmission line accurately.And the feature point detection adopting wire to customize and SIFT feature vector and the assemblage characteristic of bent wave characteristic vector vector ensure that the integrality of the accuracy of image mosaic and wire in stitching image, to avoid in traditional images splicing the problems such as wire fracture fragmentation.
The present invention can overcome transmission line of electricity fixed-wing unmanned plane and make an inspection tour the problem that Sequential images mosaic occurs compared with big error, eliminate dislocation and error hiding phenomenon, complete the quick high accuracy splicing that transmission line of electricity fixed-wing unmanned plane makes an inspection tour sequence image, realize overall the complete clear of census information of transmission line of electricity and represent.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet that power transmission line extracts;
Fig. 2 is the schematic flow sheet of image mosaic.
Embodiment
Below in conjunction with the drawings and specific embodiments, content of the present invention is described in further details.
Embodiment
Fixed-wing unmanned plane for power industry is maked an inspection tour image and is required higher to image mosaic technology, and Major Difficulties is that wire there will be dislocation, the not first-class phenomenon of docking when splicing.The present invention mainly can be divided into 4 parts: data encasement, and power transmission line extracts, the coupling of image and splicing, image registration and fusion.
The present invention is directed to the deficiency of the marginal information only utilizing image in conventional transmission lines extractive technique, extracting textural characteristics in image by having anisotropic warp wavelet, realizing the estimation to power transmission line distributed areas.Again according to the two-wire feature of power transmission line in high-definition picture, realize the further screening to the power transmission line extracted, the power transmission line completing high accuracy height robust extracts.
On the basis that power transmission line extracts, power transmission line is chosen SIFT feature point, the detection of SIFT feature point is carried out in other regions in the picture simultaneously, ensure that the SIFT feature of power transmission line is counted basically identical with other unique point numbers in image, and the feature of supplementing in traditional SIFT feature vector based on warp wavelet, finally utilize Euclidean distance to determine the unique point of coupling to complete registration and the fusion of image.The flow process that power transmission line extracts please refer to shown in Fig. 1, and the flow process that images match and splicing are merged please refer to shown in Fig. 2.Concrete technical step is as follows:
1. the preparation of data and the training of sample
1) data encasement: be the training subimage of 100 × 100 by sample cutting
2) PCA(Principal Component Analysis principal component analysis (PCA)) calculating, PCA algorithm a kind ofly conventional to process information based on variable covariances matrix, compresses and the effective ways of extracting.
A) 1 is utilized in step 2) method introduced carries out sample characteristics extraction
B) computation of mean values vector, the parameters such as projection matrix
3) RBF(Radial Basis Function radial basis function) training of neural network
A) to subimage in comprise power transmission line and be labeled as 1, what do not comprise power transmission line is labeled as-1
B) RBF network parameter, nodes and the number of plies is set
C) with the form of random weight for starting point carries out neural metwork training.(at least repeating ten times)
D) RBF neural is exported, training result.
2. Image semantic classification
Two steps are mainly divided into first to be carry out geometric calibration and image enhaucament.
1) geometric calibration: the geometric calibration utilizing unmanned plane POS information to carry out Aerial Images, corrects because fixed-wing unmanned plane makes an inspection tour distortion, the distortion of the image that shooting angle and lens distortion cause.
2) image enhaucament: realize carrying out denoising when fully retaining boundary information to image by anisotropic filtering.
3. extract based on texture information and wire dual edge unique point power transmission line
1) the power transmission line distributed areas based on warp wavelet are detected:
A) image is converted into gray level image;
B) subregion of the schema extraction image of moving window is utilized;
C) feature is extracted by warp wavelet;
2) 2 are utilized in step 1) the PCA parameter that obtains carries out Feature Dimension Reduction;
A) feature after dimensionality reduction is carried out based in step 1 3) in the differentiation of RBF neural that obtains, and to mark this region be 1;
B) 4 are repeated in step 3) c ~ f step is until image traversal is complete;
C) power transmission line distributed areas (mark value >0) are extracted according to the judgement situation of subregion.
3) extract based on power transmission line candidate edge
A) image is converted into gray level image;
B) being carried out the detection of image border by Canny operator, is original candidates edge;
C) by Probabilistic Hough Transform, complete the extraction of power transmission line, and the power transmission line of fracture is connected, preserve linear position information.
4) power transmission line based on warp wavelet and power transmission line dual edge feature screens
A) preliminary screening of power transmission line candidate edge is carried out according to the power transmission line distributed areas obtained;
B) in the candidate edge through preliminary screening, carry out based on direction, the screening of the base attributes such as length;
C) by the candidate edge of further screening, choose and to be parallel to each other and distance is no more than two straight lines of M (M<5) individual pixel, getting its average is reference line.
If the reference line d) do not satisfied condition, M=M-1 carries out standard straight-line retrieval again, until M=0 and single straight line.
E) in original candidates edge, utilize probability Hough to carry out potential power transmission line extraction, be namely strictly parallel to reference line, and parallel distance meets parallel wire separation criteria, continuity requires to reduce.
F) power transmission line extracted is exported.
4. based on SIFT image registration and the fusion of power transmission line customization
1) the SIFT feature point customized based on wire detects
1. the method chosen with SIFT feature point selects N number of unique point on the power transmission line extracted.In graphical rule space, travel through the point on all power transmission lines, judge the relation of point in itself and neighborhood, if the value of this point be greater than or less than neighborhood value a little, then this point is candidate feature point
2. other regions utilize SIFT feature point detection algorithm to choose N number of unique point in the picture, require distant with the power transmission line unique point selected in step 1) and are evenly distributed with image-region as far as possible.
2) unique point describes vector and generates: utilize SIFT operator extraction 128 D feature vectors, then extract bent wave characteristic describe vector as unique point together with SIFT feature.
3) coupling of unique point: using the Euclidean distance between descriptor as the similarity criteria of Feature Points Matching.Adopt preferential k-d to set and carry out first search to search two approximate KNN unique points of each unique point, as find out unique point p Euclidean distance recently and two time near neighbors feature point q ' and q "; then calculate p and q ' and p and q " ratio r of Euclidean distance between two group descriptors, if ratio r is less than defined threshold T, then be considered as that the match is successful, receiving station is a pair match point in image sequence to (p, q '), otherwise it fails to match.
4) image mosaic and fusion: realize the splicing between two width or multiple image by image weighting average method.
Above-listed detailed description is illustrating for possible embodiments of the present invention, and this embodiment is also not used to limit the scope of the claims of the present invention, and the equivalence that all the present invention of disengaging do is implemented or changed, and all should be contained in the scope of the claims of this case.
Claims (8)
1., for the method that transmission line of electricity fixed-wing unmanned plane tour image accurately splices, it is characterized in that, it comprises the following steps:
Step 1, data encasement and sample training; Described step 1 comprises the following steps:
Step 1.1, be the training subimage of multiple formed objects by sample cutting;
Step 1.2, the warp wavelet based on USFFT algorithm is adopted to carry out sample characteristics extraction, to obtain training sample to described training subimage;
Step 1.3, calculated the PCA parameter of described training sample by PCA algorithm, described PCA parameter comprises mean value vector and projection matrix;
The training of step 1.4, RBF neural;
Step 2, based on textural characteristics and linear feature to power transmission line feature extraction, to obtain power transmission line distributed areas and power transmission line extracts data; Extract data by described power transmission line distributed areas to described power transmission line to screen, obtain power transmission line by the dual edge feature of power transmission line and extract result;
Step 3, described power transmission line extracted to SIFT image registration and fusion that result customizes based on power transmission line;
Described step 2 comprises the following steps:
Step 2.1, textural characteristics is utilized to carry out the detection of power transmission line distributed areas;
Step 2.2, being carried out the detection of image border by Canny operator, is original candidates edge;
Step 2.3, by Probabilistic Hough Transform, straight-line detection is carried out to power transmission line, complete the extraction of power transmission line, and the power transmission line of fracture is connected, preserve linear position information;
Step 2.4, according to power transmission line distributed areas detect carry out of result to Hough straight-line detection screen, the straight line in power transmission line distributed areas is chosen as power transmission line original candidates edge;
Step 2.5, carry out the further screening at power transmission line original candidates edge according to following rule:
A) direction at original candidates edge meets the direction of the power transmission line calculated according to geography information;
B) length at original candidates edge exceedes the threshold value of specifying;
Step 2.6, by the candidate edge of further screening, choose and be parallel to each other and distance two straight lines that are M pixel, getting its average is reference line, described M=power transmission line diameter/image resolution ratio;
If the reference line do not satisfied condition, M=M-1 carries out standard straight-line retrieval again, until M=0 and single straight line;
The result that step 2.7, output power transmission line extract.
2. method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image according to claim 1, it is characterized in that, described step 1.4 comprises the following steps:
To described training subimage in include power transmission line be labeled as 1, what do not comprise power transmission line is labeled as-1;
Setting RBF network parameter, nodes and the number of plies;
With the form of random weight for starting point at least repeats ten neural metwork trainings;
Export the network weight of the RBF neural after having trained.
3. method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image according to claim 1, is characterized in that, described step 2.1 utilizes textural characteristics to carry out the detection of power transmission line distributed areas and comprises the following steps:
Step 2.1.1, Aerial Images is converted into gray level image;
Step 2.1.2, utilize the subregion of the schema extraction gray level image of moving window;
Step 2.1.3, power transmission line regional determination is carried out to every sub regions;
Step 2.1.4, travel through all subregions, complete the region detection of power transmission line distribution, wherein, area detection result intermediate value be greater than 0 be power transmission line distributed areas.
4. method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image according to claim 3, it is characterized in that, described step 2.1.3 comprises the following steps:
Step 2.1.3.1, USFFT algorithm is utilized to extract bent wave characteristic to subregion;
Step 2.1.3.2, antithetical phrase extracted region to bent wave characteristic carry out based on principal component analysis Feature Dimension Reduction;
Step 2.1.3.3, utilize and train the RBF neural that obtains to differentiate the feature after dimensionality reduction, comprise power transmission line if judge, by region corresponding for testing result all assignment be 1, otherwise assignment is-1.
5. method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image according to claim 4, it is characterized in that, described step 3 comprises the following steps:
Step 3.1, the SIFT feature point detection customized based on power transmission line;
Step 3.2, unique point describe vector and generate, and utilize SIFT operator extraction SIFT feature vector, then extract bent wave characteristic describe vector as unique point together with SIFT feature vector;
The coupling of step 3.3, unique point, the coupling of described unique point describes the similarity criteria of the Euclidean distance between vector as Feature Points Matching using unique point, adopt preferential k-d to set and carry out first search to search two approximate KNN unique points of each unique point, calculate the Euclidean distance of two corresponding with it respectively approximate KNN unique points of this unique point, and calculate the ratio of two Euclidean distances, if ratio is less than threshold value, then the match is successful, otherwise, again mate;
Step 3.4, image mosaic and fusion, realize the splicing between two width or multiple image by image weighting average method.
6. method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image according to claim 5, it is characterized in that, described step 3.1 comprises the following steps:
On the power transmission line extracted, N number of unique point is selected with SIFT feature point detection algorithm, the point on all power transmission lines is traveled through in graphical rule space, judge the relation of point in itself and neighborhood, if the value of this point be greater than or less than neighborhood value a little, then this point is candidate feature point;
Other regions utilize SIFT feature point detection algorithm to choose N number of unique point equally in the picture, require distant with described candidate feature point and are evenly distributed in image-region as far as possible.
7. method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image according to claim 1, is characterized in that, also comprise Image semantic classification between described step 1 and step 2.
8. method of accurately splicing for transmission line of electricity fixed-wing unmanned plane tour image according to claim 7, it is characterized in that, described Image semantic classification comprises the following steps:
Utilize the geometric calibration that unmanned plane POS information is carried out Aerial Images, to correct because fixed-wing unmanned plane makes an inspection tour distortion, the distortion of the image that shooting angle and lens distortion cause;
Realize carrying out denoising when fully retaining boundary information to image by anisotropic filtering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410020480.8A CN103761722B (en) | 2014-01-16 | 2014-01-16 | Fixed wing unmanned aerial vehicle touring image accurately-splicing method for power transmission line |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410020480.8A CN103761722B (en) | 2014-01-16 | 2014-01-16 | Fixed wing unmanned aerial vehicle touring image accurately-splicing method for power transmission line |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103761722A CN103761722A (en) | 2014-04-30 |
| CN103761722B true CN103761722B (en) | 2015-03-11 |
Family
ID=50528955
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410020480.8A Active CN103761722B (en) | 2014-01-16 | 2014-01-16 | Fixed wing unmanned aerial vehicle touring image accurately-splicing method for power transmission line |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103761722B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105139374A (en) * | 2015-07-09 | 2015-12-09 | 天津航天中为数据系统科技有限公司 | Power transmission line tracking method and system |
| CN105825207B (en) * | 2016-04-20 | 2019-05-14 | 北京航空航天大学 | The high-voltage line detection method and device of fragmentation |
| CN106355580B (en) * | 2016-09-22 | 2020-07-14 | 云南电网有限责任公司电力科学研究院 | Method and device for detecting toppling of tower |
| CN106595500B (en) * | 2016-11-21 | 2019-06-14 | 云南电网有限责任公司电力科学研究院 | Electric power line ice-covering thickness measurement method based on unmanned plane binocular vision |
| CN106971453B (en) * | 2017-04-06 | 2020-01-14 | 深圳怡化电脑股份有限公司 | Paper money fragment splicing method and device |
| CN106932688B (en) * | 2017-04-24 | 2023-09-05 | 国家电网公司 | Power transmission line detector and power transmission line detection system based on unmanned aerial vehicle |
| WO2018218481A1 (en) * | 2017-05-31 | 2018-12-06 | 深圳市大疆创新科技有限公司 | Neural network training method and device, computer system and mobile device |
| CN109544608B (en) * | 2018-03-22 | 2023-10-24 | 广东电网有限责任公司清远供电局 | Unmanned aerial vehicle image acquisition characteristic registration method |
| CN108377033A (en) * | 2018-04-20 | 2018-08-07 | 国网辽宁省电力有限公司沈阳供电公司 | Polling transmission line based on multi-rotor unmanned aerial vehicle compares modification system with line map |
| US10692200B2 (en) | 2018-04-25 | 2020-06-23 | Microsoft Technology Licensing, Llc | Sensor location determination for geographical region |
| CN110827335B (en) * | 2019-11-01 | 2020-10-16 | 北京推想科技有限公司 | Mammary gland image registration method and device |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102722734A (en) * | 2012-05-23 | 2012-10-10 | 南京航空航天大学 | Image target identification method based on curvelet domain bilateral two-dimension principal component analysis |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7840625B2 (en) * | 2005-04-07 | 2010-11-23 | California Institute Of Technology | Methods for performing fast discrete curvelet transforms of data |
-
2014
- 2014-01-16 CN CN201410020480.8A patent/CN103761722B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102722734A (en) * | 2012-05-23 | 2012-10-10 | 南京航空航天大学 | Image target identification method based on curvelet domain bilateral two-dimension principal component analysis |
Non-Patent Citations (4)
| Title |
|---|
| 基于PCA+RBF的多光谱卫星遥感影像分类;周峰等;《物探化探计算技术》;20080229;第30卷(第2期);第159页右栏第1段 * |
| 无人机图像自动拼接问题研究;肖波;《昆明理工大学硕士学位论文》;20131008;全文 * |
| 无人机航拍图像配准方法的研究;陈润锋;《辽宁大学硕士学位论文》;20130501;无人机航拍图像配准方法的研究 * |
| 输电线路典型目标图像识别技术研究;张少平;《南京航空航天大学硕士学位论文》;20121201;第1页第7段，第4页第2段，第5页第3-9段，第25页第3段-第26页第6段，以及图4.2 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103761722A (en) | 2014-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103761722B (en) | Fixed wing unmanned aerial vehicle touring image accurately-splicing method for power transmission line | |
| US10762376B2 (en) | Method and apparatus for detecting text | |
| Fathi et al. | Detecting road intersections from GPS traces | |
| CN108596055B (en) | Airport target detection method of high-resolution remote sensing image under complex background | |
| CN103679674B (en) | Method and system for splicing images of unmanned aircrafts in real time | |
| CN110245678B (en) | Image matching method based on heterogeneous twin region selection network | |
| Liu et al. | Automated vehicle extraction and speed determination from QuickBird satellite images | |
| CN103593832A (en) | Method for image mosaic based on feature detection operator of second order difference of Gaussian | |
| CN103500449B (en) | Visible remote sensing image cloud detection method of optic on a kind of star | |
| CN111340701A (en) | Circuit board image splicing method for screening matching points based on clustering method | |
| CN104599258A (en) | Anisotropic characteristic descriptor based image stitching method | |
| CN111914838A (en) | License plate recognition method based on text line recognition | |
| CN106548169A (en) | Fuzzy literal Enhancement Method and device based on deep neural network | |
| CN110781757A (en) | Airport pavement foreign matter identification and positioning method and system | |
| CN113159043A (en) | Feature point matching method and system based on semantic information | |
| CN103729631A (en) | Vision-based connector surface feature automatically-identifying method | |
| CN104851089A (en) | Static scene foreground segmentation method and device based on three-dimensional light field | |
| CN114419165A (en) | Camera external parameter correcting method, device, electronic equipment and storage medium | |
| Hu et al. | Automatic extraction of main road centerlines from high resolution satellite imagery using hierarchical grouping | |
| CN111213154A (en) | Lane line detection method, lane line detection equipment, mobile platform and storage medium | |
| CN110675442A (en) | Local stereo matching method and system combined with target identification technology | |
| Kampffmeyer et al. | Dense dilated convolutions merging network for semantic mapping of remote sensing images | |
| CN103578105A (en) | Method for multisource different-type image registration based on region features | |
| CN113538585B (en) | High-precision multi-target intelligent identification, positioning and tracking method and system based on unmanned aerial vehicle | |
| CN114358133A (en) | Method for detecting looped frames based on semantic-assisted binocular vision SLAM |
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 | ||
| CP01 | Change in the name or title of a patent holder |
Address after: Maintenance Test Center, Building A4, No. 181, Kexue Avenue, Science City, Luogang District, Guangzhou, Guangdong 510663 Patentee after: China Southern Power Grid Corporation Ultra High Voltage Transmission Company Electric Power Research Institute Patentee after: TIANJIN ZHONG WEI AEROSPACE DATA SYSTEM TECHNOLOGY Co.,Ltd. Address before: Maintenance Test Center, Building A4, No. 181, Kexue Avenue, Science City, Luogang District, Guangzhou, Guangdong 510663 Patentee before: MAINTENANCE & TEST CENTRE, CSG EHV POWER TRANSMISSION Co. Patentee before: TIANJIN ZHONG WEI AEROSPACE DATA SYSTEM TECHNOLOGY Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20140430 Assignee: Yidao Smart (Beijing) Supply Chain Management Co.,Ltd. Assignor: China Southern Power Grid Corporation Ultra High Voltage Transmission Company Electric Power Research Institute Contract record no.: X2023980041958 Denomination of invention: A Method for Accurate Mosaicing of Inspection Images of Fixed Wing Unmanned Aerial Vehicles on Transmission Lines Granted publication date: 20150311 License type: Common License Record date: 20230918 |
|
| EE01 | Entry into force of recordation of patent licensing contract |