CN106295652A - A kind of linear feature matching process and system - Google Patents
A kind of linear feature matching process and system Download PDFInfo
- Publication number
- CN106295652A CN106295652A CN201610600639.2A CN201610600639A CN106295652A CN 106295652 A CN106295652 A CN 106295652A CN 201610600639 A CN201610600639 A CN 201610600639A CN 106295652 A CN106295652 A CN 106295652A
- Authority
- CN
- China
- Prior art keywords
- line
- straight line
- pixel
- matching
- candidate
- 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
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to field of Computer Graphics, disclose a kind of linear feature matching process and system, by obtaining digital picture;SIFT algorithm is utilized to carry out feature point extraction and coupling described digital picture;Gaussian filtering denoising;Pixel is sorted out;Obtain matching line Candidate Set;Calculate the lap with reference to straight line with the candidate matches straight line in described matching line Candidate Set;According to rectilinear geometry characteristic and half-tone information, it is achieved matching line segments.Improve the suitability of linear feature coupling, reduce the conforming requirement to matching line attribute.
Description
Technical field
The present invention relates to field of Computer Graphics, particularly relate to a kind of linear feature matching process and system.
Background technology
Linear feature is the principal outline feature of the artificial field scapes such as building, plays in the three-dimensional model structure of building
Controlling effect.In picture, linear feature is very important middle level descriptor in picture, contrasts and puts feature, straight line
Feature more can reflect the whole geometry topological relation in picture.In overall linear feature, even if partial straight lines is blocked not
Visible, remainder stills provide the enough information reconstructing three-dimensional model for building.For building, how from
Picture extracts the linear feature of building, and the coupling realizing linear feature is the pass describing three-dimensional model building structure
Key.
Mainly consider geometric attribute and the neighborhood half-tone information of straightway at conventional Algorism of Matching Line Segments algorithm, Ru Huangliang is bright
Etc. utilize Geometrical algebra to mate line segment, wherein every straightway its midpoint vector, length, ternary of direction vector composition
Group represents;Wang Zhiheng et al. sets up straight line description by gray scale, gradient and the gradient magnitude in calculating straight line neighborhood and realizes straight
The coupling of line feature, additionally also has the matching line segments that some geometrical constraint methods adding other complete.Building in shooting
In picture owing to blocking, the impact of the factor such as illumination and extraction algorithm disappearance, coupling in different pictures may be destroyed straight
The concordance of line attribute so that the coupling of linear feature there is also a lot of problems.General task scene need to process many simultaneously
The image of type, this just requires that matching line segments algorithm can be applied to various image type.And existing matching line segments method
It is confined to the situation that the concordance of matching line attribute is higher, not there is versatility.
Summary of the invention
The present invention provides a kind of linear feature matching process and system, solves prior art cathetus characteristic matching and only limits to
In the situation that the concordance of matching line attribute is higher, not there is the technical problem of versatility.
It is an object of the invention to be achieved through the following technical solutions:
A kind of linear feature matching process, including:
Obtain digital picture;
Described digital picture utilize scale invariant feature conversion SIFT algorithm carry out feature point extraction and coupling, to obtain
Match point set;
Utilize Gaussian filter that described digital picture is filtered denoising;
Calculate gradient magnitude and the gradient direction of each pixel in described digital picture, described pixel is pressed gradient magnitude
Sequential iteration from big to small, and pixel close for gradient direction is classified as into uniform areas;
According to the gray value information of pixels all in described uniform areas, determine the center point coordinate of rectangle, wherein, rectangle
The linear feature edge that centrage is described uniform areas;
The one group of linear feature intersected with the virtual line section that Corresponding matching point is end points in acquisition left images, described one
Group linear feature is matching line Candidate Set;
According to the epipolar-line constraint relation between different visual angles image, calculate with reference in straight line and described matching line Candidate Set
The lap of candidate matches straight line;
Calculate in left image the angle phase of every straight line and the candidate matches straight line in matching line Candidate Set in right image
Like degree, selecting described angle similarity more than the candidate matches straight line presetting similarity threshold is candidate's straight line, supports according to line
Territory gray scale similarity, the candidate's straight line calculating the correlation coefficient of straightway gray average the highest is matching line.
A kind of linear feature matching system, including:
Acquisition module, is used for obtaining digital picture;
Matching module, for utilizing scale invariant feature conversion SIFT algorithm to carry out feature point extraction to described digital picture
And coupling, to obtain match point set;
Filtration module, is used for utilizing Gaussian filter that described digital picture is filtered denoising;
Classifying module, for calculating gradient magnitude and the gradient direction of each pixel in described digital picture, by described picture
Gradient magnitude sequential iteration from big to small pressed by vegetarian refreshments, and pixel close for gradient direction is classified as into uniform areas;
First computing module, for the gray value information according to pixels all in described uniform areas, determines in rectangle
Heart point coordinates, wherein, the centrage of rectangle is the linear feature edge of described uniform areas;
First matching line segments module, intersects with the virtual line section that Corresponding matching point is end points for obtaining in left images
One group of linear feature, described one group of linear feature is matching line Candidate Set;
Second computing module, for according to the epipolar-line constraint relation between different visual angles image, calculates with reference to straight line and institute
State the lap of candidate matches straight line in matching line Candidate Set;
Second matching line segments module, for calculate in left image every straight line with in matching line Candidate Set in right image
The angle similarity of candidate matches straight line, the candidate matches straight line selecting described angle similarity to be more than default similarity threshold is
Candidate's straight line, according to line supporting domain gray scale similarity, calculates candidate's straight line that the correlation coefficient of straightway gray average is the highest
For matching line.
The present invention provides a kind of linear feature matching process and system, by obtaining digital picture;To described digital picture
SIFT algorithm is utilized to carry out feature point extraction and coupling;Gaussian filtering denoising;Pixel is sorted out;Obtain matching line Candidate Set;
Calculate the lap with reference to straight line with the candidate matches straight line in described matching line Candidate Set;According to rectilinear geometry characteristic and
Half-tone information, carries out linear feature coupling.Improve the suitability of linear feature coupling, reduce to matching line attribute
The requirement of cause property.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to institute in embodiment
The accompanying drawing used is needed to be briefly described, it should be apparent that, the accompanying drawing in describing below is only some enforcements of the present invention
Example, for those of ordinary skill in the art, on the premise of not paying creative work, also can obtain according to these accompanying drawings
Obtain other accompanying drawing.
Fig. 1 is the flow chart of a kind of linear feature matching process of the embodiment of the present invention;
Fig. 2 is the structural representation of a kind of linear feature matching system of the embodiment of the present invention.
Detailed description of the invention
Understandable for enabling the above-mentioned purpose of the present invention, feature and advantage to become apparent from, real with concrete below in conjunction with the accompanying drawings
The present invention is further detailed explanation to execute mode.
As it is shown in figure 1, the flow chart of a kind of linear feature matching process provided for the embodiment of the present invention, including:
Step 101, acquisition digital picture;
Step 102, described digital picture is utilized scale invariant feature conversion SIFT algorithm carry out feature point extraction and
Join, to obtain match point set;
Step 103, utilize Gaussian filter that described digital picture is filtered denoising;
Step 104, the gradient magnitude calculating each pixel in described digital picture and gradient direction, press described pixel
Gradient magnitude sequential iteration from big to small, and pixel close for gradient direction is classified as into uniform areas;
Wherein, step 104 specifically may include that
Calculate each pixel in described digital picture gradient magnitude G (x, y) and gradient direction α, wherein,
(x y) is pixel p (x, y) place to i
Pixel value;
Described pixel is pressed gradient magnitude order from big to small, is more than the pixel of predetermined angle threshold value with gradient magnitude
As seed points;
By the way of iteration, the neighborhood territory pixel of seed points is scanned for, search the pixel that gradient direction is close, and
Described pixel is classified as into uniform areas.
Step 105, gray value information according to pixels all in described uniform areas, determine the center point coordinate of rectangle,
Wherein, the centrage of rectangle is the linear feature edge of described uniform areas;
Wherein, step 105 specifically may include that
Rectangular configuration is utilized to approach towards unified region, rectangular centre point coordinates (cx,cy) it isWherein, g (i) is the gray value of pixel i, and x (i), y (i) are the coordinates of pixel i.
The one group of straight line intersected with the virtual line section that Corresponding matching point is end points in step 106, acquisition left images is special
Levying, described one group of linear feature is matching line Candidate Set;
Step 107, according to the epipolar-line constraint relation between different visual angles image, calculate with reference to straight line and described matching line
The lap of the candidate matches straight line in Candidate Set;
Wherein, step 107 specifically may include that
Obtain the polar curve that the candidate matches straight line in described matching line Candidate Set is corresponding in left image, obtain described pole
Intersection point between line and described reference straight line;
According to described intersection point, determine the overlapping portion with reference to straight line with the candidate matches straight line in described matching line Candidate Set
Point.
Step 108, according to rectilinear geometry characteristic and half-tone information, carry out linear feature coupling;
Wherein, the angle of every straight line and the candidate matches straight line in matching line Candidate Set in right image is calculated in left image
Degree similarity, selecting described angle similarity more than the candidate matches straight line presetting similarity threshold is candidate's straight line, according to line
Supporting domain gray scale similarity, the candidate's straight line calculating the correlation coefficient of straightway gray average the highest is matching line.
Step 108 specifically may include that
Calculate every straight line in left image respective tiltedly with the candidate matches straight line in matching line Candidate Set in right image
Rate k1, k2;
Obtain the angle σ between intersecting straight lines, wherein, σ=arctanK1-arctanK2;
Straight line angle similarity is An_sim (LA,LB)=cos (σA-σB);
Wherein, σAFor left image cathetus LAWith the angle of straightway, σBFor candidate's straight line L in right imageBWith line correspondence
The angle of section.
Line supporting domain gray scale similarity.With the specific region of gray distribution features around line feature, matrix G (L) is straight line L
Gray scale Description Matrix, contain the half-tone information of structure near L.
For highlighting the importance near linear pixel, according to the vertical dimension of distance straightway, the average of each subregion
Compose a distance weighting, obtain the straightway gray average vector of Weight
Gw(L)=[w1g1w2g2 … w2r+1g2r+1]T
The correlation coefficient of the straightway gray average of Weight is estimated as grey similarity
It is more than one by calculating the angle similarity of every straight line candidate matches straight line corresponding with right image in left image
Candidate's straight line of fixed similarity threshold (such as: 0.8), calculates grey similarity and estimates the highest candidate's straight line for coupling directly
Line.
The present invention provides a kind of linear feature matching process and system, by obtaining digital picture;To described digital picture
SIFT algorithm is utilized to carry out feature point extraction and coupling;Gaussian filtering denoising;Pixel is sorted out;Obtain matching line Candidate Set;
Calculate the lap with reference to straight line with the candidate matches straight line in described matching line Candidate Set;According to rectilinear geometry characteristic and
Half-tone information, it is achieved matching line segments.Improve the suitability of linear feature coupling, reduce the concordance to matching line attribute
Requirement.
The embodiment of the present invention additionally provides a kind of linear feature matching system, as in figure 2 it is shown, include:
Acquisition module 210, is used for obtaining digital picture;
Matching module 220, for utilizing scale invariant feature conversion SIFT algorithm to carry out characteristic point to described digital picture
Extract and coupling, to obtain match point set;
Filtration module 230, is used for utilizing Gaussian filter that described digital picture is filtered denoising;
Classifying module 240, for calculating gradient magnitude and the gradient direction of each pixel in described digital picture, by described
Pixel presses gradient magnitude sequential iteration from big to small, and pixel close for gradient direction is classified as into uniform areas;
First computing module 250, for the gray value information according to pixels all in described uniform areas, determines rectangle
Center point coordinate, wherein, the centrage of rectangle is the linear feature edge of described uniform areas;
First matching line segments module 260, is used for obtaining in left images and the virtual line section that Corresponding matching point is end points
The one group of linear feature intersected, described one group of linear feature is matching line Candidate Set;
Second computing module 270, for according to the epipolar-line constraint relation between different visual angles image, calculate with reference to straight line with
The lap of the candidate matches straight line in described matching line Candidate Set;
Second matching line segments module 280, is used for calculating every straight line and matching line Candidate Set in right image in left image
In the angle similarity of candidate matches straight line, select described angle similarity straight more than the candidate matches presetting similarity threshold
Line is candidate's straight line, according to line supporting domain gray scale similarity, calculates the candidate that the correlation coefficient of straightway gray average is the highest
Straight line is matching line.
Through the above description of the embodiments, those skilled in the art is it can be understood that can be by the present invention
Software adds the mode of required hardware platform and realizes, naturally it is also possible to all implemented by hardware, but a lot of in the case of before
Person is more preferably embodiment.Based on such understanding, technical scheme background technology is contributed whole or
Person's part can embody with the form of software product, and this computer software product can be stored in storage medium, as
ROM/RAM, magnetic disc, CD etc., including some instructions with so that a computer equipment (can be personal computer, service
Device, or the network equipment etc.) perform each embodiment of the present invention or the method described in some part of embodiment.
Being described in detail the present invention above, specific case used herein is to the principle of the present invention and embodiment party
Formula is set forth, and the explanation of above example is only intended to help to understand method and the core concept thereof of the present invention;Meanwhile, right
In one of ordinary skill in the art, according to the thought of the present invention, the most all can change
Part, in sum, this specification content should not be construed as limitation of the present invention.
Claims (5)
1. a linear feature matching process, it is characterised in that including:
Obtain digital picture;
Described digital picture utilize scale invariant feature conversion SIFT algorithm carry out feature point extraction and coupling, to obtain coupling
Point set;
Utilize Gaussian filter that described digital picture is filtered denoising;
Calculate gradient magnitude and the gradient direction of each pixel in described digital picture, described pixel is pressed gradient magnitude from greatly
To little sequential iteration, and pixel close for gradient direction is classified as into uniform areas;
According to the gray value information of pixels all in described uniform areas, determine the center point coordinate of rectangle, wherein, in rectangle
Heart line is the linear feature edge of described uniform areas;
Obtain in left images the one group of linear feature intersected with the virtual line section that Corresponding matching point is end points, described one group straight
Line is characterized as matching line Candidate Set;
According to the epipolar-line constraint relation between different visual angles image, calculate with reference to straight line and the time in described matching line Candidate Set
Select the lap of matching line;
Calculate in left image the angle similarity of every straight line and the candidate matches straight line in matching line Candidate Set in right image,
Selecting described angle similarity is candidate's straight line more than the candidate matches straight line of default similarity threshold, according to line supporting domain gray scale
Similarity, the candidate's straight line calculating the correlation coefficient of straightway gray average the highest is matching line.
Linear feature matching process the most according to claim 1, it is characterised in that every in the described digital picture of described calculating
The gradient magnitude of individual pixel and gradient direction, press gradient magnitude sequential iteration from big to small by described pixel, and by gradient
The close pixel in direction is classified as into the step of uniform areas, including:
Calculate each pixel in described digital picture gradient magnitude G (x, y) and gradient direction α, wherein,
(x y) is pixel p (x, y) pixel at place to i
Value;
Described pixel is pressed gradient magnitude order from big to small, using gradient magnitude more than predetermined angle threshold value pixel as
Seed points;
By the way of iteration, the neighborhood territory pixel of seed points is scanned for, search the pixel that gradient direction is close, and by institute
State pixel and be classified as into uniform areas.
Linear feature matching process the most according to claim 1, it is characterised in that described according to institute in described uniform areas
There is the gray value information of pixel, determine the center point coordinate of rectangle, including:
Utilizing rectangular configuration to approach towards unified region, rectangular centre point coordinates (cx, cy) isWherein, g (i) is the gray value of pixel i, and x (i), y (i) are the coordinates of pixel i.
Linear feature matching process the most according to claim 1, it is characterised in that described according to different visual angles image between
Epipolar-line constraint relation, calculate the step of lap with reference to the candidate matches straight line in straight line and described matching line Candidate Set
Suddenly, including:
Obtain the polar curve that the candidate matches straight line in described matching line Candidate Set is corresponding in left image, obtain described polar curve with
Described with reference to the intersection point between straight line;
According to described intersection point, determine the lap with reference to straight line with the candidate matches straight line in described matching line Candidate Set.
5. a linear feature matching system, it is characterised in that including:
Acquisition module, is used for obtaining digital picture;
Matching module, for described digital picture utilized scale invariant feature conversion SIFT algorithm carry out feature point extraction and
Join, to obtain match point set;
Filtration module, is used for utilizing Gaussian filter that described digital picture is filtered denoising;
Classifying module, for calculating gradient magnitude and the gradient direction of each pixel in described digital picture, by described pixel
By gradient magnitude sequential iteration from big to small, and pixel close for gradient direction is classified as into uniform areas;
First computing module, for the gray value information according to pixels all in described uniform areas, determines the central point of rectangle
Coordinate, wherein, the centrage of rectangle is the linear feature edge of described uniform areas;
First matching line segments module, for obtaining intersected in left images with the virtual line section that Corresponding matching point is end points
Group linear feature, described one group of linear feature is matching line Candidate Set;
Second computing module, for according to the epipolar-line constraint relation between different visual angles image, calculating reference straight line and described
Join the lap of the candidate matches straight line that straight line candidates is concentrated;
Second matching line segments module, for calculating every straight line and the candidate in matching line Candidate Set in right image in left image
The angle similarity of matching line, selecting described angle similarity more than the candidate matches straight line presetting similarity threshold is candidate
Straight line, according to line supporting domain gray scale similarity, calculate the highest candidate's straight line of the correlation coefficient of straightway gray average for
Join straight line.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610600639.2A CN106295652A (en) | 2016-07-27 | 2016-07-27 | A kind of linear feature matching process and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610600639.2A CN106295652A (en) | 2016-07-27 | 2016-07-27 | A kind of linear feature matching process and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106295652A true CN106295652A (en) | 2017-01-04 |
Family
ID=57662511
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610600639.2A Pending CN106295652A (en) | 2016-07-27 | 2016-07-27 | A kind of linear feature matching process and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106295652A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107038721A (en) * | 2017-03-27 | 2017-08-11 | 西安交通大学 | A kind of line detection method based on LAPJV algorithms |
CN107833204A (en) * | 2017-10-23 | 2018-03-23 | 国网辽宁省电力有限公司营口供电公司 | A kind of visible detection method of the substation equipment infrared image based on topology matching |
CN109490072A (en) * | 2018-10-09 | 2019-03-19 | 广东交通职业技术学院 | A kind of civil engineering work detection system and its detection method |
CN109711321A (en) * | 2018-12-24 | 2019-05-03 | 西南交通大学 | A kind of wide Baseline Images unchanged view angle linear feature matching process of structure adaptive |
CN109902695A (en) * | 2019-03-01 | 2019-06-18 | 辽宁工程技术大学 | One kind is towards as to the correction of linear feature matched line feature and method of purification |
CN110223355A (en) * | 2019-05-15 | 2019-09-10 | 大连理工大学 | A kind of feature mark poiX matching process based on dual epipolar-line constraint |
CN110490301A (en) * | 2019-04-25 | 2019-11-22 | 中国科学院自动化研究所南京人工智能芯片创新研究院 | Line character description method based on convolutional neural networks |
CN112183596A (en) * | 2020-09-21 | 2021-01-05 | 湖北大学 | Linear segment matching method and system combining local grid constraint and geometric constraint |
CN113469167A (en) * | 2021-07-21 | 2021-10-01 | 浙江大华技术股份有限公司 | Method, device, equipment and storage medium for recognizing meter reading |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102298779A (en) * | 2011-08-16 | 2011-12-28 | 淮安盈科伟力科技有限公司 | Image registering method for panoramic assisted parking system |
CN102822864A (en) * | 2009-06-16 | 2012-12-12 | 日本电气株式会社 | Image identifier matching device |
WO2014133251A1 (en) * | 2013-02-27 | 2014-09-04 | 주식회사 케이쓰리아이 | Matching point extraction system using feature points of data inquiry result of lsh algorithm and method for same |
CN104966290A (en) * | 2015-06-12 | 2015-10-07 | 天津大学 | Self-adaptive weight three-dimensional matching method based on SIFT descriptor |
-
2016
- 2016-07-27 CN CN201610600639.2A patent/CN106295652A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102822864A (en) * | 2009-06-16 | 2012-12-12 | 日本电气株式会社 | Image identifier matching device |
CN102298779A (en) * | 2011-08-16 | 2011-12-28 | 淮安盈科伟力科技有限公司 | Image registering method for panoramic assisted parking system |
WO2014133251A1 (en) * | 2013-02-27 | 2014-09-04 | 주식회사 케이쓰리아이 | Matching point extraction system using feature points of data inquiry result of lsh algorithm and method for same |
CN104966290A (en) * | 2015-06-12 | 2015-10-07 | 天津大学 | Self-adaptive weight three-dimensional matching method based on SIFT descriptor |
Non-Patent Citations (2)
Title |
---|
梁艳: ""基于近景图像序列的建筑物三维模型重建研究"", 《中国博士学位论文全文数据库 基础科学辑》 * |
梁艳等: ""利用局部仿射不变及核线约束的近景影像直线特征匹配"", 《武汉大学学报》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107038721A (en) * | 2017-03-27 | 2017-08-11 | 西安交通大学 | A kind of line detection method based on LAPJV algorithms |
CN107038721B (en) * | 2017-03-27 | 2020-01-10 | 西安交通大学 | Linear detection method based on LAPJV algorithm |
CN107833204B (en) * | 2017-10-23 | 2021-06-01 | 国网辽宁省电力有限公司营口供电公司 | Visual detection method of infrared image of substation equipment based on topology matching |
CN107833204A (en) * | 2017-10-23 | 2018-03-23 | 国网辽宁省电力有限公司营口供电公司 | A kind of visible detection method of the substation equipment infrared image based on topology matching |
CN109490072A (en) * | 2018-10-09 | 2019-03-19 | 广东交通职业技术学院 | A kind of civil engineering work detection system and its detection method |
CN109711321A (en) * | 2018-12-24 | 2019-05-03 | 西南交通大学 | A kind of wide Baseline Images unchanged view angle linear feature matching process of structure adaptive |
CN109902695A (en) * | 2019-03-01 | 2019-06-18 | 辽宁工程技术大学 | One kind is towards as to the correction of linear feature matched line feature and method of purification |
CN109902695B (en) * | 2019-03-01 | 2022-12-20 | 辽宁工程技术大学 | Line feature correction and purification method for image pair linear feature matching |
CN110490301A (en) * | 2019-04-25 | 2019-11-22 | 中国科学院自动化研究所南京人工智能芯片创新研究院 | Line character description method based on convolutional neural networks |
CN110223355B (en) * | 2019-05-15 | 2021-01-05 | 大连理工大学 | Feature mark point matching method based on dual epipolar constraint |
CN110223355A (en) * | 2019-05-15 | 2019-09-10 | 大连理工大学 | A kind of feature mark poiX matching process based on dual epipolar-line constraint |
CN112183596A (en) * | 2020-09-21 | 2021-01-05 | 湖北大学 | Linear segment matching method and system combining local grid constraint and geometric constraint |
CN113469167A (en) * | 2021-07-21 | 2021-10-01 | 浙江大华技术股份有限公司 | Method, device, equipment and storage medium for recognizing meter reading |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106295652A (en) | A kind of linear feature matching process and system | |
Fu et al. | Removing rain from single images via a deep detail network | |
CN102360421B (en) | Face identification method and system based on video streaming | |
Matsuyama et al. | Structural analysis of natural textures by Fourier transformation | |
CN103902977B (en) | Face identification method and device based on Gabor binary patterns | |
Cai et al. | FCSR-GAN: Joint face completion and super-resolution via multi-task learning | |
CN104299260A (en) | Contact network three-dimensional reconstruction method based on SIFT and LBP point cloud registration | |
CN101996406A (en) | No-reference structural sharpness image quality evaluation method | |
CN110491132A (en) | Vehicle based on video frame picture analyzing, which is disobeyed, stops detection method and device | |
CN102682428A (en) | Fingerprint image computer automatic mending method based on direction fields | |
CN105184825A (en) | Indoor-scene-oriented mobile augmented reality method | |
CN102136065A (en) | Face super-resolution method based on convex optimization | |
Pirzada et al. | Analysis of edge detection algorithms for feature extraction in satellite images | |
CN103810729A (en) | Raster image vectorizing method based on contour line | |
CN108986024A (en) | A kind of regularly arranged processing method of laser point cloud based on grid | |
CN103544488A (en) | Face recognition method and device | |
CN105792131A (en) | Positioning method and system | |
CN104899576A (en) | Face identification feature extraction method based on Gabor conversion and HOG | |
CN106156691A (en) | The processing method of complex background image and device thereof | |
CN103914829A (en) | Method for detecting edge of noisy image | |
Klemp et al. | LDFA: Latent diffusion face anonymization for self-driving applications | |
CN104992176B (en) | A kind of Chinese character extracting method towards an inscription on a tablet | |
CN105740878B (en) | The Symbol Recognition and device of GIS-Geographic Information System | |
CN110473255A (en) | A kind of ship bollard localization method divided based on multi grid | |
CN108921147A (en) | A kind of black smoke vehicle recognition methods based on dynamic texture and transform domain space-time characteristic |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170104 |