CN111861876A - Hardware fitting automatic identification method based on shape characteristics - Google Patents
Hardware fitting automatic identification method based on shape characteristics Download PDFInfo
- Publication number
- CN111861876A CN111861876A CN202010823587.1A CN202010823587A CN111861876A CN 111861876 A CN111861876 A CN 111861876A CN 202010823587 A CN202010823587 A CN 202010823587A CN 111861876 A CN111861876 A CN 111861876A
- Authority
- CN
- China
- Prior art keywords
- shape
- centroid
- mser
- convex hull
- fitting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000010606 normalization Methods 0.000 claims description 13
- 230000005540 biological transmission Effects 0.000 abstract description 5
- 238000007689 inspection Methods 0.000 abstract description 4
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 18
- 239000011159 matrix material Substances 0.000 description 12
- 230000009466 transformation Effects 0.000 description 9
- 238000004422 calculation algorithm Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000000354 decomposition reaction Methods 0.000 description 3
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 3
- 239000010931 gold Substances 0.000 description 3
- 229910052737 gold Inorganic materials 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000009795 derivation Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/02—Affine transformations
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of power transmission line inspection, and discloses a hardware fitting automatic identification method based on shape characteristics.
Description
Technical Field
The invention relates to the technical field of power transmission line inspection, in particular to a hardware fitting automatic identification method based on shape characteristics.
Background
Because overhead lines are widely distributed and run in the open air for a long time, the overhead lines are often influenced by the change of the surrounding environment and the nature, and because the types of distribution line equipment are more and more complicated than those of a power transmission line, all parts of the power transmission line need to be observed, checked and measured so as to master the running condition of the line, discover equipment defects in time and threaten the safety of the line.
The current novel line robot patrols and examines can not receive environmental factor influence sight fixed point and patrol and examine, and although various gold utensils on the overhead transmission line all have standard shape and size, because the reason of mounted position and observation angle, the gold utensil image can change, has brought the challenge for machine automatic identification gold utensil.
The Local Affinity Frame (LAF) algorithm proposed by Stepan Obdrzalek is a feature normalization algorithm for MSER feature construction, and is commonly used in image recognition processing. The LAF algorithm firstly detects MSER characteristics in the picture, then extracts affine covariant characteristics on the MSER characteristics, and finally constructs a local affine frame by utilizing the affine covariant characteristics and normalizes the picture. However, the LAF algorithm has the defects that the calculation mode of the main direction is too complex, the description of the normalized image is too simple, and the recognition rate is not high.
Disclosure of Invention
The invention aims to provide a hardware fitting automatic identification method based on shape characteristics, and the method is used for solving the problem that a machine cannot automatically identify hardware fittings when a line robot patrols.
In order to solve the above problems, the present invention provides a hardware fitting automatic identification method based on shape characteristics, which includes the following steps:
s1: acquiring an initial picture including the electric power fitting to be identified, and extracting a picture MSER characteristic region;
s2: determining the main direction of the MSER characteristic region:
s2a, determining a convex hull and a centroid which are formed by the MSER characteristic region point set;
s2b, dividing the area surrounded by the convex hull into a plurality of triangles by utilizing the centroid and each side forming the convex hull, wherein the number of the triangles corresponds to each side forming the convex hull, and selecting the direction of the bottom side of the triangle with the largest area as the main direction;
s3: normalizing the MSER characteristic region by the main direction determined in S2 b;
s4: MSER describes:
s4a, extracting the centroid of the MSER characteristic after normalization processing in the step S3, equally dividing the MSER characteristic into 18 intervals on the circumference by taking the centroid as the center, and counting the sum of the distances from the centroid to each edge point in each interval;
s4 b: making a distance histogram from the data of the sum of the distances obtained in the step S4a as an identification shape descriptor of the electric power fitting to be identified;
s5: and matching the identification shape descriptor obtained in the step S4b with all the power fitting shape descriptors in the shape descriptor database, wherein the model of the power fitting shape descriptor successfully matched is the model of the power fitting to be identified.
Further, the number of pixels of the power fitting image area to be recognized in the above step S1 is greater than 100.
Further, the nearest neighbor method is adopted for identifying the shape descriptor and the power fitting shape descriptor matching in step S5.
Further, the threshold of the nearest neighbor method is 0.8, and only nearest neighbors smaller than the threshold are confirmed as matching features.
The invention provides a hardware fitting automatic identification method based on shape characteristics, which realizes the automatic identification of hardware fittings by calculating affine invariant shape characteristics of hardware fitting images, and particularly uses an MSER characteristic extraction algorithm to extract an MSER region containing electric power fitting images, determines a convex hull and a centroid formed by an MSER characteristic region point set to find out the main direction of the MSER characteristic region, then uses a normalization algorithm to transform the MSER characteristic region, and finally establishes an identification shape descriptor for the MSER characteristic after normalization processing and matches the identification shape descriptor with an electric power fitting shape descriptor in a database, wherein successful matching means correct identification of the electric power fittings, so that the work load of inspection personnel is reduced, the inspection quality of a robot is improved, and online real-time and continuous hardware fitting identification is realized.
Drawings
FIG. 1 is a diagram of a convex hull for a set of points in an embodiment of the invention.
Fig. 2 is a schematic diagram of the processes of calculating a convex hull of an image region, determining a main direction, and normalizing according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of an actual image MSER feature normalization process in the embodiment of the present invention.
FIG. 4 is a schematic diagram illustrating the distance and direction from the centroid to each edge point according to the embodiment of the present invention.
Fig. 5 is a schematic diagram of counting distances in each directional interval and obtaining a histogram in the embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
In the description of the present invention, it is to be understood that the terms "upper", "lower", "left", "right", "top", "bottom", and the like, indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience in describing the present invention and simplifying the description, but do not indicate or imply that the device or element referred to must have a specific orientation, be constructed in a specific orientation, and be operated, and thus, should not be construed as limiting the present invention.
With reference to fig. 1 to 4, a hardware fitting automatic identification method based on shape features is schematically shown in an embodiment of the present invention, and includes the following steps:
s1: acquiring an initial picture including the electric power fittings to be identified, wherein the initial picture is generally a picture shot through a robot pod, preferably, the number of pixels in an image area of the electric power fittings to be identified can be more than 100, and extracting a MSER characteristic area of the picture;
s2: determining the main direction of the MSER characteristic region:
s2a, determining a convex hull and a centroid which are formed by the MSER characteristic region point set;
s2b, dividing the area surrounded by the convex hull into a plurality of triangles by utilizing the centroid and each side forming the convex hull, wherein the number of the triangles corresponds to each side forming the convex hull, and selecting the direction of the bottom side of the triangle with the largest area as the main direction, thereby greatly simplifying the calculation process of the main direction and improving the calculation accuracy;
s3: normalizing the MSER characteristic region by the main direction determined in S2 b;
as an example:
the normalization process in step S3 is performed by a calculation method based on the area covariance and the principal direction, and the procedure is as follows:
given a covariance matrix of a region
And the centroid
A principal direction transformed by affine transformation of the region
(the main direction is the main direction determined in step S2 b);
the covariance matrix of the region is first subjected to Cholesky decomposition, i.e.
The main direction is changed, and the main direction is changed,
the direction information is added to the transformation matrix to obtain,
plus centroid
The coordinates of (a) get an affine transformation:
the mathematical derivation process of the normalization method based on the area covariance and the principal direction is given below:
the centroids of the regions are affine covariant, the covariance matrix of the two affine corresponding regions
And
the relationship between them can be expressed by the formula (5).
Wherein
In order to normalize the covariance matrix of the image,
is a covariance matrix of the original image,
is an affine transformation of the normalized image into the original image.
To pair
Cholesky decomposition can be performed to obtain:
wherein
Is an upper triangular matrix. For a rotational transformation
The following relationship holds:
. Equation (5) can thus be written again:
thus, it is possible to provide
Equation (6) shows the Cholesky decomposition of the covariance matrix of the region
Equivalent to affine transformation from a normalized standard coordinate system to an affine coordinate system (without taking into account the displacement of the origin of coordinates) under the condition of one rotation difference, the purpose of adding a rotation transformation in equation (3) is to supplement
The rotation of the matrix by the difference. Because the centroid of the region is affine covariant and comprises the displacement of the coordinate origin, the centroid coordinate of the region can be obtainedAffine transformation in equation (4).
Since the stability of the points is not as good as the covariance matrix of the regions, the algorithm in this embodiment normalizes the shape image using a normalization method based on the region covariance and the principal direction. The covariance matrix and centroid of the region can be directly calculated from the coordinates of the points in the region, and the difficulty is how to determine the principal direction
In the present embodiment, the main direction is determined
The method comprises the following steps:
calculating a convex hull of a given image region (MSER feature region extracted in step S1), dividing a region surrounded by the convex hull into a plurality of triangles by using the centroid and each of the edges constituting the convex hull, wherein the number of triangles corresponds to each of the edges constituting the convex hull, selecting a direction of a base of a triangle in which the area is largest as a main direction
。
S4: MSER describes:
s4a, extracting the centroid of the MSER characteristic after normalization processing in the step S3, equally dividing the MSER characteristic into 18 intervals on the circumference by taking the centroid as the center, and counting the sum of the distances from the centroid to each edge point in each interval;
s4 b: a distance histogram is created from the data of the sum of the distances obtained in S4a as an identification shape descriptor of the electric power fitting to be identified. The MSER description method comprises the steps of equally dividing MSER characteristics into 18 intervals on a circumference by taking a centroid as a center, then counting distance information from the centroid to each edge contour point, distributing the distance information to each angle interval, adding distance values in each angle interval to form a distance histogram distributed according to the angle intervals, and finally normalizing the distance histogram to eliminate the influence of shape dimension. The shape description algorithm can accurately describe the shape information, can tolerate the shape information loss to a certain degree, has higher accuracy of shape matching and has better real-time performance.
The detailed construction process of the shape descriptor in this embodiment is as shown in fig. 4 and 5, and the distance from the centroid C of the normalized MSER feature to each edge point is calculated in fig. 4, and the direction of the connecting line between the centroid and the edge point is calculated. In fig. 4, the 360-degree angle is uniformly divided into 18 intervals, the distances from the centroid to the edge points are distributed into 18 angular regions according to the direction of the connecting line between the centroid and the edge points, the sum of the distances in each interval is calculated, and the sum of the distances in each interval is made into a distance histogram shown in fig. 5.
S5: and matching the identification shape descriptor obtained in the step S4b with all the power fitting shape descriptors in the shape descriptor database, wherein the model of the power fitting shape descriptor successfully matched is the model of the power fitting to be identified.
Preferably, the nearest neighbor method is used to identify the shape descriptor and the power fitting shape descriptor in step S5.
Preferably, the threshold of the nearest neighbor method is 0.8, and only nearest neighbors smaller than the threshold are confirmed as matching features, so that mismatching can be effectively reduced, and the influence on the number of matching features is small.
The threshold value of the nearest neighbor method in this embodiment is defined as: the proportion between the nearest neighbor distance and the next nearest neighbor distance is used for ensuring the matching accuracy and avoiding mismatching. For example: the similarity between shape descriptors a and B is 0.8 and the similarity between shape descriptors a and C is 0.78 (the second largest similarity), but the difference between 0.8 and 0.78 is not much, and it is possible that a and C are actually more similar, and to avoid this, a and B are considered to be matched only when the similarity between shape descriptors a and B is 0.8 and the similarity between shape descriptors a and C is 0.5, 0.5:0.8 < 0.8 (the threshold value of the ratio between nearest neighbor distance and next nearest neighbor distance).
The principle and process of determining the main direction are as follows:
the shape image of any MSER feature region is a set of pixel points in a plane and has a unique convex hull, and as shown in fig. 1, the edges of the convex hull are formed by the straight portion of the edge curve of the image region and the double tangent of the concave portion, so that the convex hull contains the direction information, and the embodiment will extract the main direction by the convex hull of the shape image.
According to the method for extracting the main direction of the convex hull, as shown in a in FIG. 2, the black area is two shape images corresponding to affine, and a polygon enclosed outside the black area is the convex hull of the shape images. The polygon forming the convex hull contains a plurality of line segments (i.e., each side of the polygon) from which an affine covariant line segment needs to be selected. From the knowledge of high geometry: the area ratio of the two polygons is an affine invariant. From this it can be inferred that: a region is divided into a plurality of blocks, and after the affine transformation of the region, the block with the largest area is also the largest in the blocks of the region. The centroid of the region and each side of the convex hull are used for dividing the region surrounded by the convex hull into a plurality of triangles, the triangle with the largest area is selected, the area of the triangle is also the largest after the convex hull is subjected to affine transformation, therefore, the base edge of the triangle is affine covariant, and the direction of the base edge can be used as the main direction. As shown in b in fig. 2, the position marked by the cross line, i.e., the centroid of the shape image, is connected to one of the edges by taking the centroid as a vertex, and when the enclosed triangle is the triangle with the largest area in the convex hull, the direction of the base thereof is the main direction, and c in fig. 2 shows the corresponding images of the two shape images in a in fig. 2 after normalization processing according to the main direction.
Fig. 3 shows the complete MSER feature normalization process under the actual image, the left picture shows an ellipse containing the MSER feature and its covariance matrix, the middle picture determines the main direction by calculating the centroid and convex hull of the MSER feature, and the right picture is the normalization result.
It should be understood that the terms "first", "second", etc. are used herein to describe various information, but the information should not be limited to these terms, which are only used to distinguish one type of information from another. For example, "first" information may also be referred to as "second" information, and similarly, "second" information may also be referred to as "first" information, without departing from the scope of the present invention.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and substitutions can be made without departing from the technical principle of the present invention, and these modifications and substitutions should also be regarded as the protection scope of the present invention.
Claims (4)
1. A hardware fitting automatic identification method based on shape characteristics is characterized by comprising the following steps:
s1: acquiring an initial picture including the electric power fitting to be identified, and extracting a picture MSER characteristic region;
s2: determining the main direction of the MSER characteristic region:
s2a, determining a convex hull and a centroid which are formed by the MSER characteristic region point set;
s2b, dividing the area surrounded by the convex hull into a plurality of triangles by utilizing the centroid and each side forming the convex hull, wherein the number of the triangles corresponds to each side forming the convex hull, and selecting the direction of the bottom side of the triangle with the largest area as the main direction;
s3: normalizing the MSER characteristic region by the main direction determined in S2 b;
s4: MSER describes:
s4a, extracting the centroid of the MSER characteristic after normalization processing in the step S3, equally dividing the MSER characteristic into 18 intervals on the circumference by taking the centroid as the center, and counting the sum of the distances from the centroid to each edge point in each interval;
s4 b: making a distance histogram from the data of the sum of the distances obtained in the step S4a as an identification shape descriptor of the electric power fitting to be identified;
s5: and matching the identification shape descriptor obtained in the step S4b with all the power fitting shape descriptors in the shape descriptor database, wherein the model of the power fitting shape descriptor successfully matched is the model of the power fitting to be identified.
2. The method for automatically identifying hardware fittings based on shape characteristics as claimed in claim 1, wherein the number of pixels of the image area of the electric power fittings to be identified in the step S1 is more than 100.
3. The automatic shape feature-based hardware fitting identification method according to claim 1, wherein the identification of the shape descriptor in step S5 to match the power fitting shape descriptor uses a nearest neighbor method.
4. The automatic shape feature-based hardware recognition method according to claim 3, wherein the threshold of the nearest neighbor method is 0.8, and only nearest neighbors smaller than the threshold are confirmed as matching features.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010823587.1A CN111861876B (en) | 2020-08-17 | 2020-08-17 | Automatic hardware fitting identification method based on shape characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010823587.1A CN111861876B (en) | 2020-08-17 | 2020-08-17 | Automatic hardware fitting identification method based on shape characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111861876A true CN111861876A (en) | 2020-10-30 |
| CN111861876B CN111861876B (en) | 2024-01-26 |
Family
ID=72970180
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010823587.1A Active CN111861876B (en) | 2020-08-17 | 2020-08-17 | Automatic hardware fitting identification method based on shape characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111861876B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113112487A (en) * | 2021-04-20 | 2021-07-13 | 广东电网有限责任公司 | Method, device, equipment and storage medium for identifying thermal defects |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090238460A1 (en) * | 2006-04-28 | 2009-09-24 | Ryuji Funayama | Robust interest point detector and descriptor |
| CN103400384A (en) * | 2013-07-22 | 2013-11-20 | 西安电子科技大学 | Large viewing angle image matching method capable of combining region matching and point matching |
| CN104867126A (en) * | 2014-02-25 | 2015-08-26 | 西安电子科技大学 | Method for registering synthetic aperture radar image with change area based on point pair constraint and Delaunay |
| CN105354578A (en) * | 2015-10-27 | 2016-02-24 | 安徽大学 | Multi-target object image matching method |
| CN110956186A (en) * | 2019-11-21 | 2020-04-03 | 山东浪潮通软信息科技有限公司 | Image recognition method, device and medium |
-
2020
- 2020-08-17 CN CN202010823587.1A patent/CN111861876B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090238460A1 (en) * | 2006-04-28 | 2009-09-24 | Ryuji Funayama | Robust interest point detector and descriptor |
| CN103400384A (en) * | 2013-07-22 | 2013-11-20 | 西安电子科技大学 | Large viewing angle image matching method capable of combining region matching and point matching |
| CN104867126A (en) * | 2014-02-25 | 2015-08-26 | 西安电子科技大学 | Method for registering synthetic aperture radar image with change area based on point pair constraint and Delaunay |
| CN105354578A (en) * | 2015-10-27 | 2016-02-24 | 安徽大学 | Multi-target object image matching method |
| CN110956186A (en) * | 2019-11-21 | 2020-04-03 | 山东浪潮通软信息科技有限公司 | Image recognition method, device and medium |
Non-Patent Citations (3)
| Title |
|---|
| 杨燕: "复杂场景图像特征分析与匹配方法研究", 中国博士学位论文全文数据库 信息科技 * |
| 梁艳;盛业华;杨林;: "大交角近景影像的仿射不变特征匹配方法研究", 测绘科学, no. 02 * |
| 祁曦;陈志云;: "结合SIFT和Delaunay三角网的遥感图像配准算法", 计算机系统应用, no. 10 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113112487A (en) * | 2021-04-20 | 2021-07-13 | 广东电网有限责任公司 | Method, device, equipment and storage medium for identifying thermal defects |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111861876B (en) | 2024-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20190362178A1 (en) | Object Symmetry Axis Detection Method Based on RGB-D Camera | |
| Kawaguchi et al. | Detection of eyes from human faces by Hough transform and separability filter | |
| Soltany et al. | Fast and accurate pupil positioning algorithm using circular Hough transform and gray projection | |
| US20050129293A1 (en) | Fingerprint minutiae matching using scoring techniques | |
| WO2019190405A1 (en) | Method and apparatus for detecting condition of a bolt on a bolted structure | |
| CN110097021B (en) | MTCNN-based face pose estimation method | |
| US20200211220A1 (en) | Method for Identifying an Object Instance and/or Orientation of an Object | |
| CN103727930A (en) | Edge-matching-based relative pose calibration method of laser range finder and camera | |
| CN109886124A (en) | One kind describing the matched texture-free metal parts grasping means of subgraph based on harness | |
| CN109389625B (en) | Three-dimensional image registration method based on multi-scale descriptor screening and mismatching | |
| CN111402224A (en) | Target identification method for power equipment | |
| CN105608448A (en) | LBP characteristic extraction method based on face key points and LBP characteristic extraction device based on face key points | |
| CN109447036A (en) | A kind of segmentation of image digitization and recognition methods and system | |
| CN108090486A (en) | Image processing method and device in a kind of game of billiards | |
| CN113850815B (en) | Workpiece point cloud obtaining method and device, electronic equipment and storage medium | |
| CN108280471B (en) | Machine vision-based change-over switch state identification method | |
| US6735343B2 (en) | Polygon finder and pruned tree geometric match method | |
| CN111861876B (en) | Automatic hardware fitting identification method based on shape characteristics | |
| CN110472651B (en) | Target matching and positioning method based on edge point local characteristic value | |
| CN114333023A (en) | Face gait multi-mode weighting fusion identity recognition method and system based on angle estimation | |
| CN113269234B (en) | Connecting piece assembly detection method and system based on target detection | |
| CN110633749A (en) | Three-dimensional point cloud identification method based on improved viewpoint feature histogram | |
| CN114022441A (en) | Defect detection method for irregular hardware | |
| CN116363177A (en) | Three-dimensional point cloud registration method based on geometric transformation and Gaussian mixture model | |
| Glachet et al. | Finding the perspective projection of an axis of revolution |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |