CN107228860B - Gear defect detection method based on image rotation period characteristics - Google Patents
Gear defect detection method based on image rotation period characteristics Download PDFInfo
- Publication number
- CN107228860B CN107228860B CN201710511442.6A CN201710511442A CN107228860B CN 107228860 B CN107228860 B CN 107228860B CN 201710511442 A CN201710511442 A CN 201710511442A CN 107228860 B CN107228860 B CN 107228860B
- Authority
- CN
- China
- Prior art keywords
- image
- gear
- rotation
- detection
- defect
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Abstract
The invention discloses a gear defect detection method based on image rotation period characteristics, which comprises a detection pretreatment stage and a defect detection stage; the detection preprocessing stage comprises two processes of noise elimination and image morphology operation; the detection stage comprises the following steps: J1) carrying out image binarization operation; J2) extracting an image contour; J3) determining a target region of interest; J4) image refilling noise prevention; J5) gear tooth number determination; J6) gear image rotation; J7) acquiring a mean template image; J8) defect localization based on image differences. The invention utilizes the computer vision technology to preprocess the gear original image, utilizes the gear image rotation operation to obtain the template image, and matches the template image to detect the defects. The method has the characteristics of high detection precision, strong robustness and high speed.
Description
Technical Field
The invention relates to the field of gear appearance defect detection, in particular to a gear appearance defect rapid detection method based on image rotation periodicity characteristics.
Background
Computer Integrated Manufacturing (CIM) technology is a product of the combination of information technology development and industrial production, and is a technology that promotes unification, integration, and high efficiency of production. In "high and new technology research and development plan" that was implemented in 1986 in China, CIMS is an important subject, and application of CIMS was promoted at that time, and information technology began to emerge from the corner in the industrial field more frequently with popularization of CIMS (computer/modern integrated manufacturing system). In the machining process of mechanical parts, the occurrence of defective products is inevitable due to materials, processes, or other uncontrollable factors. Therefore, defect identification of industrial products is necessary, the production efficiency can be greatly improved, the burden of manual detection is reduced, and the labor cost of production is reduced. And the labor cost is one of the major expenses of enterprises which will continue to grow in the future.
With the development of computer vision technology in recent years, the application of the technology in industry is more extensive. For a target image, the related computer vision technology can often suppress a noise signal thereof, enhance a required signal, and realize processing and further analysis of the image. The existing defect detection scheme based on image information usually matches a picture to be detected with a standard template picture, and a separate standard template image needs to be additionally provided. However, in practice, the standard template image additionally and separately acquired is different from the image to be detected in external environments such as the acquired illumination and the acquired focal length, so that the conditions of underexposure or overexposure and the like often occur, and the shooting angles are difficult to be completely consistent. This causes a large distortion and color difference between the image of the component and the image of the standard template even if the component is intact, which affects the accuracy of defect detection. In addition, the method based on additional template matching usually needs a large amount of computation time for searching and traversing in multiple dimensions of scale, rotation and translation, and often cannot meet the real-time requirement of industrial detection.
Disclosure of Invention
The invention aims to solve the problems and provides a gear defect detection method based on image rotation period characteristics. The method comprises the steps of preprocessing an original image of the gear, acquiring an image of a self-adaptive template by utilizing gear image rotation operation, and detecting defects by utilizing image difference operation and template matching. The invention has high detection precision and high speed, does not need additional standard templates, has no overhigh requirement on the image acquisition environment, and can meet the requirement of large-scale production.
The invention relates to a gear defect detection method based on image rotation period characteristics, which comprises two stages of detection pretreatment and defect detection;
the detection preprocessing stage comprises two processes of image noise elimination and morphological operation; the original image noise elimination refers to the image processing operation of an original image by using a filtering method such as Gaussian filtering and a binarization segmentation method based on a threshold value, and the image morphology operation mainly comprises the detection and feature analysis of edges and contours.
The defect detection stage comprises the following steps:
step J1) carrying out binarization operation on the threshold value of the image to be detected, and carrying out pixel-level segmentation on the object and the background in the image through the threshold value operation;
the image threshold value binarization operation refers to performing image binarization operation by using a fixed threshold value or an adaptive threshold value method.
Step J2), extracting the contour of the image to be detected to obtain the geometric contour of the target in the image area;
step J3), combining the obtained minimum outsourcing contour to determine a target interested area, and extracting a circle center area of the part as the interested area of a processing target;
the determination of the target region of interest refers to extracting a circle center region of the part by combining the minimum outsourcing contour as a processed target region of interest, and screening and distinguishing the region of interest, wherein the specific process comprises the following steps:
m1) image smoothing the contour point set;
m2) searching a minimum enclosing circle, and searching and storing the minimum enclosing circles of all contours in the image;
m3) storing radius data of the minimum circumscribed circle;
m4) represents the minimum representation of the region of interest using the largest circle radius in the minimum circumscribing circle, the circumscribing rectangle of which is chosen to expand the region of interest.
Step J4) carrying out segmentation void refilling on the gear target area in the image, and further enhancing the image rotation period characteristic;
step J5) determining the number of teeth of the outer profile of the gear based on a convex hull detection method, and further obtaining the minimum rotation period angle of the gear;
the determination of the gear tooth number refers to the determination of a certain degree of the convex part of the gear by using a convex hull detection method, and the specific process comprises the following steps:
n1) acquiring a target image;
n2) obtaining the result of the direct convex hull detection;
n3) determining a zoom factor;
the determination of the scaling factor refers to reducing the pixel points by scaling, reducing the number of detected convex hulls and obtaining the functional relation between the number of the convex hulls and the scaling factor.
N4) detection analysis of the cyclic convex hull;
n5) determining the number mutation points of the convex hulls and inducing to determine the number of the gear teeth;
determining the discontinuity refers to obtaining the discontinuity at which the number of the convex hulls changes most at each zoom level, thereby determining the number of gear teeth.
Step J6) rotating the gear image by utilizing the rotation periodicity of the gear appearance to generate a plurality of new images;
the gear image is rotated by utilizing the rotation periodicity of the gear appearance, and a plurality of new images are generated. A specific implementation may refer to constructing an image rotation function through an affine transformation or other image rotation manner, where the specific process includes the following steps:
l1) acquisition of rotation matrix:
the acquisition of the rotation transformation matrix refers to acquiring the rotation matrix through a given circle center and a given rotation angle, wherein the circle center is from the circle center corresponding to the second position of the radius sequence in the contour circumscribed circle, and the rotation angle is the gear tooth number.
L2) geometrically transform the image using a rotation matrix:
l3) correcting the size of the graph according to the rotation result:
in step L3), the area after the image rotation becomes large, and the size of the region of interest is fixed, so that the problem of image loss occurs, and the problem of deformation of the rotated image after the affine transformation is dealt with by enlarging the length and width of the rotation and correcting the rotation by the radian of the rotation.
Step J7) carrying out rotation processing by utilizing the rotation periodicity of the image and obtaining an online adaptive template image;
the acquisition of the online adaptive template image refers to that the noise point becomes light after the mean value by utilizing the symmetry of the target part, and then the light part is filtered by threshold operation to obtain a smooth image. The specific process comprises the generation of a rotation image and the generation of an adaptive template based on the rotation result.
Step J8) defect detection is carried out based on image difference operation and template matching, and difference operation is carried out on the image to detect a defect area.
The difference of the images and the positioning of the defects refer to comparing the images of the mean value of the original images based on the image difference, and further determining inconsistent parts.
The invention has the advantages that:
(1) the method does not depend on an additional standard template picture, and is very concise and convenient in practical application;
(2) a matching template is automatically generated for each gear part image sample, so that the method can be self-adaptive to different illumination environments, and the debugging cost is low;
(3) the rotation period characteristic is utilized to carry out matching positioning, the search space is small, and the operation efficiency is high;
(4) the defect type is not specially required, and the applicability to various defect types is wide.
Drawings
FIG. 1 is a schematic overall flow diagram of the present invention;
FIG. 2 is a schematic flow chart illustrating the process of determining a target region of interest according to the present invention;
FIG. 3 is a schematic flow chart of the present invention for determining gear tooth number using convex hull detection;
FIG. 4 is a schematic diagram of an original image of a gear to be detected in the present invention;
FIG. 5 is a schematic diagram of the effect of binarization processing of the gear image to be detected in the invention;
FIG. 6 is a schematic diagram of the detection effect of the unqualified gear.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
The invention discloses a gear defect detection method based on image rotation period characteristics, which comprises two stages of detection pretreatment and defect detection as shown in figure 1. In the detection preprocessing stage, preprocessing steps such as image filtering, noise elimination, binarization, morphological operation and the like are included. In the defect detection stage, the method comprises the steps of image segmentation, target contour extraction, target region detection, target region completion, contour gear number estimation, gear image rotation, online adaptive template image generation, differential matching of an original image and an obtained template image and the like.
And in the detection preprocessing stage, methods such as Gaussian filtering, morphological operation and the like are adopted to eliminate noise and enhance the gear contour signal.
Fig. 4 is a schematic diagram of an original image of a gear to be detected. The embodiment uses two-dimensional Gaussian filtering to eliminate noise which is subject to normal distribution, and the principle is that the noise is in an image (x)0,y0) The pixel value of the pixel point to be filtered is the pixel value A of the pixel point with the surrounding coordinates of (x, y)x,yCarrying out weighted accumulation replacement, wherein the specific form is as follows:
wherein the content of the first and second substances,
and respectively represent the bandwidth parameters of the Gaussian kernel in the x-axis direction and the y-axis direction of the image.
After the two-dimensional Gaussian filtering operation, bilateral filtering processing is carried out on the image to smooth edges. As shown in fig. 5, the result of binarization is saved as one of the image copies to be used, on the basis of which morphological operations are performed. In this example, the morphological operations are performed as follows: first, K times of expansion, closing operation and black cap operation are carried out, and then J times of corrosion, opening operation and top cap operation are carried out. The K and J are determined empirically from the actual data, and generally K or J is selected to be about 5-10.
The defect detection stage comprises the following steps:
step J1) carrying out binarization operation on the threshold value of the image to be detected, and carrying out pixel-level segmentation on the object and the background in the image through the threshold value operation;
the threshold binarization operation is to perform pixel-level segmentation based on the difference between an object and a background in an image, and the present embodiment performs the binarization operation by using a global fixed threshold method.
Step J2), extracting the contour of the image to be detected to obtain the geometric contour of the target in the image area;
and performing polygon fitting on the extracted contour edge point set to obtain a minimum circumscribed circle or a minimum circumscribed rectangle of the contour. The specific process comprises the steps of searching, storing, screening and drawing the outline.
Step J3), combining the obtained minimum outsourcing contour to determine a target interested area, and extracting a circle center area of the part as the interested area of a processing target;
in order to eliminate noise outside the gear range, the circle center region of the part is taken as a target region of interest for processing by combining with the minimum outsourcing outline acquisition frame, all extracted outsourcing circles are stored in an array, and then screening and distinguishing are carried out, as shown in fig. 2, the method specifically comprises the following steps:
m1) smoothing the contour point set before performing minimum outsourcing on the extracted contour to form a minimum outsourcing circle or a minimum outsourcing rectangle, thereby reducing errors and improving efficiency;
m2) finding a function of the smallest enclosing circle according to the quasi-circular shape of the part;
m3) storing the obtained minimum circumscribed circle radius;
m4) represents the minimum representation of the region of interest using the largest circle radius in the minimum circumscribing circle, the circumscribing rectangle of which is chosen to expand the region of interest.
Step J4) carrying out segmentation void refilling on the gear target area in the image, and further enhancing the image rotation period characteristic;
and (4) refilling segmentation holes in the gear target area in the image, further enhancing the rotation period characteristic of the image, and preparing for the geometric transformation operation to be performed later.
Step J5) determining the number of teeth of the outer profile of the gear based on a convex hull detection method, and further obtaining the minimum rotation period angle of the gear;
the determination of the gear tooth number, the determination of the protruding part of the gear to a certain extent is carried out by using a convex hull detection method, as shown in fig. 3, the specific process comprises the following steps:
n1) acquiring a target image;
n2) obtaining the result of the direct convex hull detection;
n3) determining the zoom factor;
and (4) determining the scaling factor by reducing the pixel points by scaling, reducing the number of the detected convex hulls and obtaining the functional relation between the number of the convex hulls and the scaling factor.
N4) detection analysis of cyclic convex hulls
N5) determining the number mutation points of the convex hulls and inducing to determine the number of the gear teeth;
and determining the number of the convex hulls from the catastrophe points, and summarizing and determining the gear tooth number, wherein the catastrophe point with the largest change of the number of the convex hulls at each zoom level is obtained, so that the gear tooth number is determined.
Step J6) rotating the gear image by utilizing the rotation periodicity of the gear appearance to generate a plurality of new images;
the method comprises the following steps of constructing a rotation function through affine transformation and mathematical rotation change of corresponding points:
l1) acquisition of rotation matrix: the circle center is from the circle center of the circumscribed circle of the inner contour in the previous step of the acquisition of the circumscribed circle of the contour, the rotation angle is the determined gear tooth number N, the angle is anticlockwise, otherwise, the angle is clockwise, and the result of the obtained rotation matrix is as follows:
wherein α represents the angle of rotation.
L2) geometrically transform the image using a rotation matrix:
l3) correcting the size of the graph according to the rotation result:
step J7) carrying out rotation processing by utilizing the rotation periodicity of the image and obtaining an online adaptive template image;
the method comprises the steps of obtaining an adaptive template image by utilizing the rotation periodicity of a gear target, namely the rotation invariance, enabling noise points to be lightened after the structure of a rotation image is equalized, and then filtering the lightened part by threshold operation to obtain a smooth image. The specific process comprises the following steps:
the method comprises the steps of storing and obtaining a rotating image, namely continuously performing the rotating operation of the previous step, rotating the image of an interested area for a plurality of times, and storing a result image selected each time; the weighted accumulated sum of the result maps is divided by the number of rotated images and averaged or otherwise smoothed to obtain the desired template image.
Step J8) defect detection is performed based on image difference operation and template matching, difference operation is performed on the image, and a defect area is detected, with the detection result shown in fig. 6.
After obtaining the required mean template image, the original image and the mean image are compared, the inconsistent part is determined by using an image difference method, the first three areas of the connected domain area are selected for frame selection, the relative position coordinates are respectively marked, and finally, the original image is marked according to the obtained coordinates to determine the detected defect area, as shown in fig. 6.
Claims (5)
1. A gear defect detection method based on image rotation period characteristics comprises two stages of detection pretreatment and defect detection;
the detection preprocessing stage comprises image noise elimination and morphological operation;
the defect detection stage comprises the following steps:
step J1) carrying out binarization operation on the threshold value of the image to be detected, and carrying out pixel-level segmentation on the object and the background in the image through the threshold value operation;
step J2), extracting the contour of the image to be detected to obtain the geometric contour of the target in the image area;
step J3), combining the obtained minimum outsourcing contour to determine a target interested area, and extracting a circle center area of the part as the interested area of a processing target;
step J4) carrying out segmentation hole refilling on the gear target area in the image;
step J5) determining the number of teeth of the outer profile of the gear based on a convex hull detection method, and further obtaining the minimum rotation period angle of the gear;
step J6) rotating the gear image by utilizing the rotation periodicity of the gear appearance to generate a plurality of new images;
step J7) carrying out rotation processing by utilizing the rotation periodicity of the image and obtaining an online adaptive template image;
step J8) defect detection is carried out based on image difference operation and template matching, and difference operation is carried out on the image to detect a defect area.
2. The method for detecting the gear defect based on the image rotation period characteristic according to claim 1, wherein in the detection preprocessing stage, the image noise elimination specifically comprises: filtering the original image by using linear filtering and threshold binary segmentation; the image morphology operation specifically comprises: detection and feature analysis of image edges and contours.
3. The method for detecting the gear defect based on the image rotation period characteristic as claimed in claim 1, wherein the step J3 specifically comprises the following steps:
m1) image smoothing the contour point set;
m2) searching a minimum enclosing circle, and searching and storing the minimum enclosing circles of all contours in the image;
m3) storing radius data of the minimum circumscribed circle;
m4) represents the minimum representation of the region of interest using the largest circle radius in the minimum circumscribing circle, the circumscribing rectangle of which is chosen to expand the region of interest.
4. The method for detecting the gear defect based on the image rotation period characteristic as claimed in claim 1, wherein the step J5 specifically comprises the following steps:
n1) acquiring a target image;
n2) obtaining the result of the direct convex hull detection;
n3) determining a zoom factor;
reducing pixel points by utilizing scaling, reducing the number of detected convex hulls, and obtaining a functional relation between the number of the convex hulls and a scaling multiple;
n4) detection analysis of the cyclic convex hull;
n5) determining the number mutation points of the convex hulls and inducing to determine the number of the gear teeth;
and obtaining the catastrophe point with the largest change of the number of the convex-concave packets at each zoom level, and determining the number of the gear teeth.
5. The method for detecting the gear defect based on the image rotation period characteristic as claimed in claim 1, wherein the step J6 specifically comprises the following steps:
l1) acquisition of rotation matrix;
the acquisition of the rotation transformation matrix refers to acquiring the rotation matrix through a given circle center and a given rotation angle, wherein the circle center is from the circle center corresponding to the second position of the radius sequence in the contour circumscribed circle, and the rotation angle is the gear tooth number;
l2) geometrically transforming the image with a rotation matrix;
l3) correcting the size of the figure according to the rotation result;
the length and width of the rotation are enlarged, and the radian of the rotation is used for correction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710511442.6A CN107228860B (en) | 2017-06-28 | 2017-06-28 | Gear defect detection method based on image rotation period characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710511442.6A CN107228860B (en) | 2017-06-28 | 2017-06-28 | Gear defect detection method based on image rotation period characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107228860A CN107228860A (en) | 2017-10-03 |
| CN107228860B true CN107228860B (en) | 2020-04-24 |
Family
ID=59935652
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710511442.6A Active CN107228860B (en) | 2017-06-28 | 2017-06-28 | Gear defect detection method based on image rotation period characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107228860B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108731606A (en) * | 2018-05-17 | 2018-11-02 | 北京化工大学 | A kind of gear parameter rapid detection method based on convexity defect key point |
| CN109064463B (en) * | 2018-08-06 | 2020-10-20 | 中车青岛四方机车车辆股份有限公司 | Method and system for measuring tooth surface pitting area rate |
| CN109472769A (en) * | 2018-09-26 | 2019-03-15 | 成都数之联科技有限公司 | A kind of bad image defect detection method and system |
| CN109461141A (en) * | 2018-10-10 | 2019-03-12 | 重庆大学 | A kind of workpiece starved detection method |
| CN110633886B (en) * | 2019-04-01 | 2020-07-31 | 中道新职坊科技发展有限公司 | Block chain-based network data sending platform, method and storage medium |
| CN110084302B (en) * | 2019-04-26 | 2022-11-08 | 西安因诺航空科技有限公司 | Crack detection method based on remote sensing image |
| CN110211113B (en) * | 2019-06-03 | 2021-09-07 | 湖南瑞智健科技有限公司 | Detection algorithm and calculation equipment for groove abnormity |
| CN110220917B (en) * | 2019-06-11 | 2021-09-07 | 江苏农林职业技术学院 | Crown cap surface defect online detection method based on image processing |
| CN110322395B (en) * | 2019-06-18 | 2023-04-14 | 电子科技大学中山学院 | Part outline shape detection method and device based on image processing and affine transformation |
| CN111028202B (en) * | 2019-11-14 | 2021-07-13 | 苏州斯玛维科技有限公司 | X-ray bubble defect image processing method, device, storage medium and system for LED chip |
| CN111353981B (en) * | 2020-02-28 | 2023-03-14 | 广州大学 | Gear detection method and system based on machine vision and storage medium |
| CN112991338B (en) * | 2021-04-27 | 2021-07-27 | 湖南大捷智能装备有限公司 | Defect detection method and device for laser cutting part |
| TWI792351B (en) * | 2021-06-15 | 2023-02-11 | 鴻海精密工業股份有限公司 | Method for detecting defects, electronic device and storage medium |
| CN115035107B (en) * | 2022-08-10 | 2022-11-08 | 山东正阳机械股份有限公司 | Axle gear working error detection method based on image processing |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1468371A (en) * | 2000-09-10 | 2004-01-14 | �²�̩������˾ | Reduction of error alarm in PCB detection |
| WO2008123459A1 (en) * | 2007-03-30 | 2008-10-16 | Shibaura Mechatronics Corporation | Apparatus and method for inspecting edge of semiconductor wafer |
| CN103578087A (en) * | 2013-11-15 | 2014-02-12 | 重庆医科大学 | Retinal image optimizing method |
| KR20140059621A (en) * | 2012-11-08 | 2014-05-16 | 주식회사 쓰리디산업영상 | A method and an apparatus for deciphering flaw of a fuse |
| CN103901044A (en) * | 2012-12-28 | 2014-07-02 | 鸿富锦精密工业（深圳）有限公司 | Detection system |
| CN105069790A (en) * | 2015-08-06 | 2015-11-18 | 潍坊学院 | Rapid imaging detection method for gear appearance defect |
| CN105841613A (en) * | 2016-03-31 | 2016-08-10 | 浙江工业大学 | Device of utilizing planar mirror set to determine qualification of shelf crossbeams |
| CN106596589A (en) * | 2016-12-24 | 2017-04-26 | 大连日佳电子有限公司 | Tray IC detection device |
-
2017
- 2017-06-28 CN CN201710511442.6A patent/CN107228860B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1468371A (en) * | 2000-09-10 | 2004-01-14 | �²�̩������˾ | Reduction of error alarm in PCB detection |
| WO2008123459A1 (en) * | 2007-03-30 | 2008-10-16 | Shibaura Mechatronics Corporation | Apparatus and method for inspecting edge of semiconductor wafer |
| KR20140059621A (en) * | 2012-11-08 | 2014-05-16 | 주식회사 쓰리디산업영상 | A method and an apparatus for deciphering flaw of a fuse |
| CN103901044A (en) * | 2012-12-28 | 2014-07-02 | 鸿富锦精密工业（深圳）有限公司 | Detection system |
| CN103578087A (en) * | 2013-11-15 | 2014-02-12 | 重庆医科大学 | Retinal image optimizing method |
| CN105069790A (en) * | 2015-08-06 | 2015-11-18 | 潍坊学院 | Rapid imaging detection method for gear appearance defect |
| CN105841613A (en) * | 2016-03-31 | 2016-08-10 | 浙江工业大学 | Device of utilizing planar mirror set to determine qualification of shelf crossbeams |
| CN106596589A (en) * | 2016-12-24 | 2017-04-26 | 大连日佳电子有限公司 | Tray IC detection device |
Non-Patent Citations (4)
| Title |
|---|
| "基于对比度与最小凸包的显著性区域检测算法";范敏 等;《仪器仪表学报》;20141210;第2328-2334页 * |
| "基于机器视觉的齿轮产品外观缺陷检测";杨淑莹 等;《天津大学学报》;20071202;第1111-1114页 * |
| "基于机器视觉的齿轮缺陷研究";王丽;《电子测试》;20160922;第181-182 * |
| 基于机器视觉齿轮微观缺陷检测技术研究;蔺志敏;《中国优秀硕士学位论文全文数据库》;20160401;第51-68页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107228860A (en) | 2017-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107228860B (en) | Gear defect detection method based on image rotation period characteristics | |
| CN108898610B (en) | Object contour extraction method based on mask-RCNN | |
| CN110826416B (en) | Bathroom ceramic surface defect detection method and device based on deep learning | |
| CN108520274B (en) | High-reflectivity surface defect detection method based on image processing and neural network classification | |
| CN109580630B (en) | Visual inspection method for defects of mechanical parts | |
| CN108364311B (en) | Automatic positioning method for metal part and terminal equipment | |
| CN107808161B (en) | Underwater target identification method based on optical vision | |
| CN113450307A (en) | Product edge defect detection method | |
| WO2017181724A1 (en) | Inspection method and system for missing electronic component | |
| CN109767445B (en) | High-precision PCB defect intelligent detection method | |
| CN111833306A (en) | Defect detection method and model training method for defect detection | |
| CN112734761B (en) | Industrial product image boundary contour extraction method | |
| CN111754538B (en) | Threshold segmentation method for USB surface defect detection | |
| JP2021168143A (en) | System and method for efficiently scoring probe in image by vision system | |
| CN111027538A (en) | Container detection method based on instance segmentation model | |
| CN113705564B (en) | Pointer type instrument identification reading method | |
| Zhou et al. | An adaptive clustering method detecting the surface defects on linear guide rails | |
| CN109544513A (en) | A kind of steel pipe end surface defect extraction knowledge method for distinguishing | |
| CN113516619A (en) | Product surface flaw identification method based on image processing technology | |
| CN116433978A (en) | Automatic generation and automatic labeling method and device for high-quality flaw image | |
| CN109064420B (en) | Multilayer PCB section image process parameter extraction method | |
| CN115880305A (en) | Visual detection method and device for surface defects of shaft parts | |
| CN113378847B (en) | Character segmentation method, system, computer device and storage medium | |
| CN113284158B (en) | Image edge extraction method and system based on structural constraint clustering | |
| CN109949245B (en) | Cross laser detection positioning method and device, storage medium and computer equipment |
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 |