CN112884746A - Character defect intelligent detection algorithm based on edge shape matching - Google Patents
Character defect intelligent detection algorithm based on edge shape matching Download PDFInfo
- Publication number
- CN112884746A CN112884746A CN202110215572.1A CN202110215572A CN112884746A CN 112884746 A CN112884746 A CN 112884746A CN 202110215572 A CN202110215572 A CN 202110215572A CN 112884746 A CN112884746 A CN 112884746A
- Authority
- CN
- China
- Prior art keywords
- image
- point
- edge
- defect
- contour
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 63
- 238000001514 detection method Methods 0.000 title claims abstract description 42
- 238000012216 screening Methods 0.000 claims abstract description 17
- 230000011218 segmentation Effects 0.000 claims abstract description 7
- 238000001914 filtration Methods 0.000 claims abstract description 4
- 238000005259 measurement Methods 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 15
- 238000004519 manufacturing process Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 230000002401 inhibitory effect Effects 0.000 claims 1
- 230000002950 deficient Effects 0.000 abstract description 5
- 238000005286 illumination Methods 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000011897 real-time detection Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000005477 standard model Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30144—Printing quality
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses an intelligent character defect detection algorithm based on edge shape matching, which comprises the following steps: the algorithm determines an optimal edge segmentation threshold value based on gradient data obtained after Sobel operator filtering according to the length of an edge connected domain, and obtains the gradient direction and coordinates of edge contour points as matching information. And according to the matching degree of the contour points, searching for the defective points by using a mode of combining rough screening and fine screening, and then marking the character defective areas by using a mode of searching for connected domains based on defective point filling. In order to enable the matching algorithm to meet the real-time requirement, a self-adaptive image pyramid algorithm and a search strategy which is ended in advance through threshold judgment are adopted. The similarity measurement adopted by the invention is not influenced by the conditions of shielding, chaos, linear and nonlinear illumination change and the like, so that the character printing defects can still be quickly and accurately positioned in the illumination change environment.
Description
Technical Field
The invention relates to character quality defect detection of electronic product printing, in particular to an intelligent character defect detection algorithm based on edge shape matching, which is suitable for detecting related printed character defects in various electronic devices.
Background
With the rapid development of industrial automation technology, the production and manufacturing level is gradually improved, and the requirements of consumers on the product quality are also continuously improved. Therefore, it is very important to control the quality of the product production process.
In many industrial productions, the traditional manual detection is still adopted in the printing quality detection link of the product, the manual detection mainly comprises the steps of carefully observing the characters, characters and numbers printed on the product through human eyes, and connecting the characters, characters and numbers with an existing experience or standard model for judgment to complete the defect detection of the product. The traditional manual detection method is flexible and can quickly adapt to the responsible detection task; however, the spirit is highly concentrated for a long time during detection, and the detection accuracy is reduced due to the influence of fatigue or psychological factors, and meanwhile, the human eye detection needs to be observed sample by sample, so that the efficiency is low, and the industrial automation requirement is difficult to meet.
Most of the existing product printing quality defect detection methods adopt a template matching algorithm based on gray level to position a printing area, and utilize the difference between a template image and the positioned area to realize the defect detection of a target.
Disclosure of Invention
In order to avoid the defects of the traditional manual detection and the defects of the existing method, an intelligent character defect detection algorithm based on edge shape matching is provided. The method is based on gradient data obtained after Sobel operator filtering, and determines an optimal edge segmentation threshold value according to the length of an edge connected domain. Obtaining the gradient direction and coordinates of the edge contour points as matching information, searching for the defective points by using a mode of combining coarse screening and fine screening according to the matching degree of the contour points, and then marking character defect areas by using a mode of searching for connected domains based on the filling of the neighborhood of the defective points. In order to enable the matching algorithm to meet the real-time requirement, a self-adaptive image pyramid algorithm and a search strategy which is ended in advance through threshold judgment are adopted. The result shows that the method can realize high-precision real-time detection of the defects of the printed characters.
In order to solve the technical problems, the invention adopts the following technical scheme:
an intelligent character defect detection algorithm based on edge shape matching comprises the following steps:
A. manufacturing a template: and extracting contour points for optimally describing the printed characters, removing the influence of textures and noise, and calculating the coordinate position of the contour points and the normalized gradient direction vector by using the obtained contour points to manufacture a template. Because the target angle in the image to be detected is random, the template is manufactured once at regular intervals of rotation angle step length in order to obtain the accurate position of the target.
B. Contour matching: and solving the gradient direction of the image to be detected, taking the point multiplication value of the gradient of the edge outline point of the template and the gradient of the corresponding point of the image to be detected as similarity measurement, and calculating the position of the target character on the image to be detected through pixel-by-pixel sliding windows.
C. Defect detection and positioning: after the position of a target on an image to be detected is obtained, according to the matching degree of the character contour points, the defect points are searched in a mode of combining coarse screening and fine screening, and the defect positions are located by a method of solving the minimum circumscribed rectangle based on the filling of the defect points.
D. And (3) accelerating a search strategy: in order to enable the defect detection algorithm to meet the real-time requirement of industrial detection, a self-adaptive image pyramid algorithm and a search strategy which is ended in advance through threshold judgment are adopted.
The character defect intelligent detection algorithm, wherein the step a specifically further comprises: firstly, a Sobel operator is adopted to obtain image gray values of edge detection in the horizontal direction and the vertical direction, but because the image edge obtained by the Sobel operator is influenced by noise, all edge gradient amplitudes are determined on the basis of image edge extraction by the Sobel operator in order to remove the influence of noise and texture, edges larger than an edge gradient threshold value are extracted from a template image and are restrained by non-maximum values, then the edges are further segmented at a high curvature point, a Two-Pass connected region marking algorithm is adopted to process the edge image to obtain the number of connected domains, and the total length of the edges is calculated to be divided by the number of the connected domains to obtain the average length of the edges. And setting the gradient amplitude corresponding to the maximum value of the average length as an edge segmentation threshold. And regarding the determination of the template rotation angle step, calculating the rotation angle of the contour point farthest from the center of the template around the center of the template according to the cosine law to calculate the rotation angle step. In order to avoid mismatching, the step length of the rotation angle is not set too large. Meanwhile, in order to avoid repeated matching, the step length of the rotation angle is not set too small. The method uses an angle corresponding to the connection line between the contour point farthest from the center of the template and the center point of the template and the rotation of 2 pixels as the rotation step length.
The character defect intelligent detection algorithm, wherein the step B specifically further comprises: after the optimal edge contour segmentation threshold is obtained, a template is manufactured by using the coordinate position of the obtained contour point and the normalized gradient direction vector, and the template is converted into a series of character contour points Pi＝(Xi，Yi)TI is 1, 2, … n, and a normalized gradient direction vector d is obtained for each pointi＝(ti，ui)T. For the image to be detected, calculating the gradient direction vector e of each point (X, Y) after normalization by using Sobel operator filteringx，y＝(vx，y，wx，y)T. The similarity metric function is:
the character defect intelligent detection algorithm, wherein the step C specifically further comprises: after contour matching, each contour point has a degree of match. In the defect extraction process, a mode of combining rough screening and fine screening is adopted to search for defect points, and defects are extracted based on a method of filling the defect points to obtain minimum circumscribed rectangles.
The intelligent character defect detection algorithm specifically comprises the step D of finishing calculation in advance at a non-target point and performing matching at the next position because the matching degree score of the region outside the target character of the image to be detected is low. Therefore, the algorithm processing speed can be effectively accelerated, and the calculated amount in the matching process is reduced. The strategy of terminating the calculation by setting the threshold value can reduce the running time of the algorithm, but does not change the complexity of the algorithm. The complexity of the algorithm can be reduced by constructing a multi-layered image pyramid.
The invention has the following advantages:
according to the intelligent character defect detection algorithm based on edge shape matching, an optimal edge segmentation threshold value is obtained according to the average length of the edge, and contour points capable of optimally describing characters are extracted, so that the influence of texture and noise is removed. And calculating the coordinate position of the contour points and the normalized gradient direction vector by using the obtained contour points to manufacture a template. Therefore, after the target is matched on the diagram to be detected, defect analysis is performed according to the matching degree of each character contour point. Using the gradient direction as a matching feature, rather than the gradient magnitude, makes the feature robust to illumination variations. And a self-adaptive image pyramid algorithm and a search strategy of judging the early termination by a threshold value are adopted, so that the method meets the real-time requirement of industrial detection. The defect position can be accurately detected in real time by targets with random angles in the image to be detected in a complex illumination environment.
Drawings
FIG. 1 is a schematic flow chart of an intelligent character defect detection algorithm based on edge shape matching according to the present invention.
FIG. 2 is a schematic diagram of a template graph and an extracted outline of a character.
Fig. 3 is a schematic diagram of rotation step determination.
FIG. 4 is a diagram of an example of a character defect image of a printed matter and a detection result.
Detailed Description
The invention is described in further detail below with reference to the following figures and detailed description:
the invention provides an intelligent character defect detection algorithm based on edge shape matching, which comprises the following specific steps as shown in figure 1:
firstly, extracting contour points for optimally describing printed characters, removing the influence of textures and noise, and solving the coordinate position of the contour points and a normalized gradient direction vector by using the obtained contour points to manufacture a template. Because the target angle in the image to be detected is random, the template is manufactured once at regular intervals of rotation angle step length in order to obtain the accurate position of the target.
For determining the rotation angle step, as shown in fig. 3, assuming that the point O is the template center point, the point a is the farthest contour point, the point B is the rotated point of the point a, the side length AB is 1, and OA is OB, and the included angle between OA and OB is determined by the cosine law
And (4) an angle.
And secondly, solving the gradient direction of the image to be detected, taking the point multiplication value of the gradient of the edge outline point of the template and the gradient of the corresponding point of the image to be detected as similarity measurement, and calculating the position of the target character on the image to be detected by pixel-by-pixel sliding window. The similarity metric function is:
in the formula (d)i＝(ti，ui)TAs template character contour points Pi＝(Xi，Yi)TI is the normalized gradient direction vector of 1, 2, … n, ex，y＝(vx，y，wx，y)TAnd normalizing the gradient direction vector of each point (X, Y) of the image to be detected.
And thirdly, after the position of the target on the image to be detected is obtained, searching for a defect point through a mode of combining coarse screening and fine screening according to the matching degree of the character contour point, and positioning the defect position based on a method of filling the defect point to obtain a minimum circumscribed rectangle.
After the target character position is detected on the image to be detected, points with the score of more than or equal to 0.75 of the contour points of the target character are removed, and points with the score of less than 0.75 are reserved. Considering the influence of noise, the contour points obtained by coarse screening may not be true defect points, and therefore further screening is needed. Considering that noise points generally exist in isolation, traversal is performed in a certain region of points obtained by coarse screening, and if the region has more than a certain number of points with scores less than 0.75, the points are considered as true defect points. After the actual defect point is obtained, it is inconvenient to determine the specific defect position because the defect point exists discretely. Therefore, in order to locate the area where the character defect is located, the gray value in the 12 × 12 area around the defect point is filled to 255, and the gray values at other positions are set to 0. Then, the minimum circumscribed rectangle where the contour of the binary image is located is found, and the minimum circumscribed rectangle is the character defect area.
And fourthly, in order to enable the defect detection algorithm to meet the real-time requirement of industrial detection, a self-adaptive image pyramid algorithm and a search strategy which is ended in advance through threshold judgment are adopted.
The matching degree score of the region outside the target character of the image to be detected is low, so a certain strategy can be adopted, the calculation is finished in advance at the non-target point, and the matching at the next position is carried out. According to the similarity metric function, the part and formula of the similarity metric can be obtained as follows:
because of the vector diAnd ex，yIs a normalized gradient direction vector, so that the sum of the first j terms is less than j/n, and the sum of the remaining n-j terms is less than (n-j)/n-1-j/n, as can be seen from the formula, so that when the partial sum satisfies Sj＜Smin-1+j/n(SminA set threshold) the calculation of the position can be ended and the calculation of the next position can be performed.
The method adopts a self-adaptive image pyramid algorithm, when the number of contour points of an image pyramid at a certain layer is about a certain number and the higher layer is smaller than the value, the character image still has more complete contour characteristic points at the layer, and the layer is determined as the highest layer of the image pyramid. The method sets the number of contour points to 100.
FIG. 4 is a diagram of an example of a character defect image of a printed matter and a detection result, and it can be seen that the contour points of the target character are accurately matched and the defect position is also accurately detected, and the detection time is within 1s, which shows that the method of the present invention can realize high-precision real-time detection of the printing defect.
It should be understood, however, that the description herein of specific embodiments is not intended to limit the invention to the particular forms disclosed, but on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the appended claims.
Claims (4)
1. An intelligent character defect detection algorithm based on edge shape matching is characterized by comprising the following steps:
A. manufacturing a template: extracting contour points for optimally describing printed characters, removing the influence of textures and noise, solving the coordinate position of the contour points and a normalized gradient direction vector by using the obtained contour points to manufacture a template, wherein the template is manufactured at intervals of a certain rotation angle step length in order to obtain the accurate position of a target because the target angle in an image to be detected is random;
B. contour matching: solving the gradient direction of the image to be detected, taking the point multiplication value of the gradient of the edge contour point of the template and the gradient of the corresponding point of the image to be detected as similarity measurement, and calculating the position of the target character on the image to be detected by pixel-by-pixel sliding window;
C. defect detection and positioning: after the position of a target on an image to be detected is obtained, searching for a defect point in a mode of combining coarse screening and fine screening according to the matching degree of the character contour point, and positioning the defect position based on a method of filling the defect point to obtain a minimum circumscribed rectangle;
D. and (3) accelerating a search strategy: in order to enable the defect detection algorithm to meet the real-time requirement of industrial detection, a self-adaptive image pyramid algorithm and a search strategy which is ended in advance through threshold judgment are adopted.
2. The intelligent character defect detection algorithm based on edge shape matching according to claim 1, wherein in the step a, the specific extraction method is as follows:
in order to remove the influence of noise and texture, determining all edge gradient amplitudes on the basis of extracting image edges by a Sobel operator, extracting edges larger than an edge gradient threshold value from a template image, inhibiting the edges through a non-maximum value, further segmenting the edges at a high curvature point, processing the edge image by adopting a Two-Pass connected region marking algorithm to obtain the number of connected domains, and calculating the total length of the edges to divide the number of the connected domains to obtain the average length of the edges. And setting the gradient amplitude corresponding to the maximum value of the average length as an edge segmentation threshold.
3. The intelligent character defect detection algorithm based on edge shape matching according to claim 1, wherein in the step B, the specific extraction method is as follows:
after the optimal edge contour segmentation threshold is obtained, a template is manufactured by using the coordinate position of the obtained contour point and the normalized gradient direction vector, and the template is converted into a series of character contour points Pi＝(Xi，Yi)TI is 1, 2, … n, and a normalized gradient direction vector d is obtained for each pointi＝(ti，ui)T. For the image to be detected, calculating the gradient direction vector e of each point (X, Y) after normalization by using Sobel operator filteringx，y＝(vx，y，wx，y)T。
4. The intelligent character defect detection algorithm based on edge shape matching according to claim 1, wherein in the step C, the specific extraction method is as follows:
after contour matching, each contour point has a degree of match. In the defect extraction process, a mode of combining rough screening and fine screening is adopted to search for defect points, and defects are extracted based on a method of filling the defect points to obtain minimum circumscribed rectangles.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110215572.1A CN112884746B (en) | 2021-02-22 | 2021-02-22 | Character defect intelligent detection algorithm based on edge shape matching |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110215572.1A CN112884746B (en) | 2021-02-22 | 2021-02-22 | Character defect intelligent detection algorithm based on edge shape matching |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112884746A true CN112884746A (en) | 2021-06-01 |
| CN112884746B CN112884746B (en) | 2024-08-27 |
Family
ID=76054608
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110215572.1A Active CN112884746B (en) | 2021-02-22 | 2021-02-22 | Character defect intelligent detection algorithm based on edge shape matching |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112884746B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114998922A (en) * | 2022-07-29 | 2022-09-02 | 成都薯片科技有限公司 | Electronic contract generation method based on format template |
| CN115050017A (en) * | 2022-08-16 | 2022-09-13 | 南通万瑞精密机械加工有限公司 | Electric power fitting surface inscription identification method based on machine vision |
| CN116452580A (en) * | 2023-06-13 | 2023-07-18 | 山东古天电子科技有限公司 | Notebook appearance quality detection method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102654902A (en) * | 2012-01-16 | 2012-09-05 | 江南大学 | Contour vector feature-based embedded real-time image matching method |
| WO2019041590A1 (en) * | 2017-08-31 | 2019-03-07 | 中国科学院微电子研究所 | Edge detection method using arbitrary angle |
| CN110220917A (en) * | 2019-06-11 | 2019-09-10 | 江苏农林职业技术学院 | A kind of crown plug surface defect online test method based on image procossing |
-
2021
- 2021-02-22 CN CN202110215572.1A patent/CN112884746B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102654902A (en) * | 2012-01-16 | 2012-09-05 | 江南大学 | Contour vector feature-based embedded real-time image matching method |
| WO2019041590A1 (en) * | 2017-08-31 | 2019-03-07 | 中国科学院微电子研究所 | Edge detection method using arbitrary angle |
| CN110220917A (en) * | 2019-06-11 | 2019-09-10 | 江苏农林职业技术学院 | A kind of crown plug surface defect online test method based on image procossing |
Non-Patent Citations (1)
| Title |
|---|
| 张代林;陈文广;谢经明;陈幼平;: "基于形状模板匹配的印刷品缺陷检测", 机械与电子, no. 12, 24 December 2013 (2013-12-24) * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114998922A (en) * | 2022-07-29 | 2022-09-02 | 成都薯片科技有限公司 | Electronic contract generation method based on format template |
| CN114998922B (en) * | 2022-07-29 | 2022-11-04 | 成都薯片科技有限公司 | Electronic contract generating method based on format template |
| CN115050017A (en) * | 2022-08-16 | 2022-09-13 | 南通万瑞精密机械加工有限公司 | Electric power fitting surface inscription identification method based on machine vision |
| CN116452580A (en) * | 2023-06-13 | 2023-07-18 | 山东古天电子科技有限公司 | Notebook appearance quality detection method |
| CN116452580B (en) * | 2023-06-13 | 2023-09-01 | 山东古天电子科技有限公司 | Notebook appearance quality detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112884746B (en) | 2024-08-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107808378B (en) | Method for detecting potential defects of complex-structure casting based on vertical longitudinal and transverse line profile features | |
| CN112884746A (en) | Character defect intelligent detection algorithm based on edge shape matching | |
| CN112819772B (en) | High-precision rapid pattern detection and recognition method | |
| CN107610111B (en) | deep learning-based welding spot image detection method | |
| CN104732536A (en) | Sub-pixel edge detection method based on improved morphology | |
| CN111862037A (en) | Method and system for detecting geometric characteristics of precision hole type part based on machine vision | |
| CN109658391B (en) | Circle radius measuring method based on contour merging and convex hull fitting | |
| CN112037203A (en) | Side surface defect detection method and system based on complex workpiece outer contour registration | |
| CN102982545B (en) | A kind of image depth estimation method | |
| CN115294099B (en) | Method and system for detecting hairline defect in steel plate rolling process | |
| CN112233116B (en) | Concave-convex mark visual detection method based on neighborhood decision and gray level co-occurrence matrix description | |
| CN112734761B (en) | Industrial product image boundary contour extraction method | |
| CN111401449B (en) | Image matching method based on machine vision | |
| CN108802051B (en) | System and method for detecting bubble and crease defects of linear circuit of flexible IC substrate | |
| CN112381140B (en) | Abrasive particle image machine learning identification method based on new characteristic parameters | |
| CN113781413B (en) | Electrolytic capacitor positioning method based on Hough gradient method | |
| CN113673396B (en) | Spore germination rate calculation method, device and storage medium | |
| CN112102189B (en) | Line structure light bar center line extraction method | |
| CN115619799B (en) | Grain image segmentation method and system based on transfer learning | |
| CN112489052A (en) | Line structure light central line extraction method under complex environment | |
| CN110853097A (en) | Biscuit identification and positioning method applied to biscuit surface pattern printing equipment | |
| CN108985294B (en) | Method, device and equipment for positioning tire mold picture and storage medium | |
| CN116188408A (en) | Canny-based vamp edge detection and identification method | |
| CN113643290B (en) | Straw counting method and device based on image processing and storage medium | |
| CN114418975A (en) | Cutting method of large-area animal fur contour image based on radius increment constraint |
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 |