CN117408995B - Power adapter appearance quality detection method based on multi-feature fusion - Google Patents
Power adapter appearance quality detection method based on multi-feature fusion Download PDFInfo
- Publication number
- CN117408995B CN117408995B CN202311687400.XA CN202311687400A CN117408995B CN 117408995 B CN117408995 B CN 117408995B CN 202311687400 A CN202311687400 A CN 202311687400A CN 117408995 B CN117408995 B CN 117408995B
- Authority
- CN
- China
- Prior art keywords
- value
- pixel point
- target pixel
- gray
- target
- 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
- 230000004927 fusion Effects 0.000 title claims abstract description 20
- 238000001514 detection method Methods 0.000 title description 9
- 230000007547 defect Effects 0.000 claims abstract description 108
- 230000011218 segmentation Effects 0.000 claims abstract description 105
- 238000000034 method Methods 0.000 claims abstract description 59
- 239000013598 vector Substances 0.000 claims abstract description 38
- 238000012216 screening Methods 0.000 claims abstract description 6
- 238000011156 evaluation Methods 0.000 claims description 17
- 238000004458 analytical method Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 abstract description 2
- 230000002950 deficient Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 239000000126 substance Substances 0.000 description 2
- 239000000758 substrate Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
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
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image segmentation, in particular to a method for detecting appearance quality of a power adapter based on multi-feature fusion. The method comprises the steps of obtaining a gray image of a power adapter; screening out target pixel points according to the gray value and the gradient amplitude value, and obtaining target feature vectors of the target pixel points; obtaining a similar characteristic value according to the target characteristic vector and the gray value of the target pixel point; and obtaining possible defect values according to the gray value, gradient and similar characteristic values of the target pixel point, the length of a target interval where the target pixel point is positioned and the width of a gray image, further obtaining an initial segmentation threshold value, determining an optimal segmentation threshold value through iterative threshold segmentation, and detecting the quality of the appearance of the power adapter. According to the invention, the initial segmentation threshold value is accurately obtained, and then the optimal segmentation threshold value is accurately determined to accurately segment the scratch defect in the gray level image through iterative threshold segmentation, so that the quality of the appearance of the power adapter is accurately detected.
Description
Technical Field
The invention relates to the technical field of image segmentation, in particular to a method for detecting appearance quality of a power adapter based on multi-feature fusion.
Background
Along with popularization and popularization of electronic equipment, each electronic equipment is provided with a power adapter, and huge safety accidents can be possibly caused by using the power adapter with potential safety hazards, for example, when the appearance of the power adapter is defective, conditions such as exposure, electric leakage and displacement of components in the power adapter are easily caused, and further safety accidents are caused, so that the electronic equipment is very important for detecting the appearance quality of the power adapter, and the safety and reliability of using the power adapter by the electronic equipment can be effectively ensured.
In the existing method, the defects of the appearance of the power adapter are detected by adopting a threshold segmentation algorithm, and because the pixel values of the scratch defects on the surface of the power adapter are very close to the pixel values of the textures of the power adapter, the setting of an initial segmentation threshold in iterative threshold segmentation is very easy to be inaccurate, and the effect of iterative threshold segmentation is determined by the initial segmentation threshold, so that the result of segmentation of the scratch defects in a gray level image is inaccurate due to the inaccurate initial segmentation threshold, and the quality of the appearance of the power adapter cannot be detected accurately.
Disclosure of Invention
In order to solve the technical problems that the initial segmentation threshold value in iterative threshold segmentation is inaccurate, so that the segmentation result of scratch defects in a gray image is inaccurate, and the quality of the appearance of a power adapter cannot be accurately detected, the invention aims to provide a multi-feature fusion-based power adapter appearance quality detection method, which adopts the following specific technical scheme:
the invention provides a multi-feature fusion-based power adapter appearance quality detection method, which comprises the following steps:
Acquiring a gray image of a power adapter;
Screening out target pixel points according to the gray value and the gradient amplitude of each pixel point in the gray image; acquiring a target feature vector of each target pixel point according to the gradient and the gray value of each target pixel point; constructing a preset window for each target pixel point, and acquiring a similar characteristic value of each target pixel point and a corresponding neighborhood target pixel point according to a target characteristic vector and a gray value of the target pixel point and the neighborhood target pixel point in each preset window;
Acquiring at least one target area according to the position distribution of the target pixel points; obtaining a possible defect value of each target pixel point according to the gray value, gradient and the similar characteristic value of each target pixel point, the length of a target interval where each target pixel point is positioned and the width of a gray image;
Acquiring an initial segmentation threshold according to the possible defect value and the gray value of each target pixel point; and according to the initial segmentation threshold, obtaining an optimal segmentation threshold through iterative threshold segmentation, and detecting the quality of the appearance of the power adapter.
Further, the method for obtaining the target pixel point comprises the following steps:
For any pixel point, when the gray value of the pixel point is greater than or equal to a preset first gray value threshold value and the gradient amplitude is greater than or equal to a preset first gradient amplitude threshold value, the pixel point is a target pixel point.
Further, the method for obtaining the target feature vector of each target pixel point according to the gradient and the gray value of each target pixel point comprises the following steps:
For any target pixel point, acquiring an included angle between the gradient direction of the target pixel point and a preset direction as a first included angle of the target pixel point;
the result of normalizing the gray value of the target pixel point is taken as a first result;
The result of normalizing the gradient amplitude of the target pixel point is taken as a second result;
the result of normalizing the first included angle of the target pixel point is used as a third result;
Taking the product of the first preset weight and the first result as a first value;
taking the product of the second preset weight and the second result as a second value;
taking the product of the third preset weight and the third result as a third value;
And the first value represents the x value in the three-dimensional coordinate system, the second value represents the y value in the three-dimensional coordinate system, the third value represents the z value in the three-dimensional coordinate system, and the target feature vector of the target pixel point in the three-dimensional coordinate system is obtained.
Further, the method for constructing a preset window for each target pixel point, and obtaining the similar characteristic value of each target pixel point and the corresponding neighborhood target pixel point according to the target characteristic vector and the gray value of the target pixel point and the neighborhood target pixel point in each preset window comprises the following steps:
Constructing a preset window by taking each target pixel point as a center, and acquiring the similarity between the central target pixel point of each preset window and the target feature vector of each neighborhood target pixel point through cosine similarity;
And obtaining similar characteristic values of the central target pixel point and the corresponding neighborhood target pixel points according to the similarity and the gray value difference of the central target pixel point of each preset window and each neighborhood target pixel point.
Further, the calculation formula of the similar characteristic value is as follows:
In the method, in the process of the invention, The similar characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point; /(I)The number of the neighbor target pixel points in the preset window of the ith target pixel point is set; /(I)The similarity between the ith target pixel point and the target feature vector of the (y) th neighborhood target pixel point in a preset window of the ith target pixel point is obtained; /(I)The gray value of the ith target pixel point; /(I)The gray value of the target pixel point in the y neighborhood in the preset window of the i target pixel point is obtained; /(I)Is a preset constant, and is larger than 0; /(I)As a function of absolute value.
Further, the method for acquiring the target area comprises the following steps:
Taking each target pixel point as a seed point, and acquiring a target region through a region growing algorithm; wherein, the stopping condition of the region growing algorithm is as follows: the gray value difference between two adjacent target pixel points is larger than a preset second gray value threshold value.
Further, the calculation formula of the possible defect value is as follows:
In the method, in the process of the invention, The possible defect value of the ith target pixel point; /(I)The first included angle is the first included angle of the ith target pixel point; The gray value of the ith target pixel point; /(I) The length of the target interval where the ith target pixel point is located; d is the width of the gray scale image; /(I)The similar characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point; exp is an exponential function based on a natural constant; norm is a normalization function; /(I)As a function of absolute value.
Further, the method for obtaining the length of the target interval comprises the following steps: for any target area, acquiring a circumscribed rectangle of the target area, determining the side length of the circumscribed rectangle, and taking the maximum side length as the length of the target area;
The width of the gray image refers to the side length corresponding to the texture in the gray image.
Further, the method for acquiring the initial segmentation threshold value comprises the following steps:
acquiring the average value of possible defect values of each target pixel point to be used as a defect distinguishing threshold;
when the possible defect value is larger than the defect distinguishing threshold value, the corresponding target pixel point is used as a first class pixel point;
when the possible defect value is smaller than or equal to the defect distinguishing threshold value, the corresponding target pixel point is used as a second-class pixel point;
Acquiring a gray value average value and a defect possible value average value of first class pixel points, and respectively serving as a first gray characteristic value and a first defect characteristic value;
taking the product of the first gray characteristic value and the first defect characteristic value as a first characteristic value;
Acquiring a gray value average value and a defect possible value average value of the second class of pixel points to respectively serve as a second gray characteristic value and a second defect characteristic value;
Taking the product of the second gray characteristic value and the second defect characteristic value as a second characteristic value;
and taking the addition result of the first characteristic value and the second characteristic value as an initial segmentation threshold value.
Further, the method for obtaining the optimal segmentation threshold value and detecting the quality of the appearance of the power adapter through iterative threshold segmentation according to the initial segmentation threshold value comprises the following steps:
Based on the initial segmentation threshold, obtaining a segmentation threshold after each iteration through iterative threshold segmentation;
Setting the pixel value of the pixel point with the gray value larger than the segmentation threshold value in the gray image as 1, and setting the pixel value of the pixel point with the gray value smaller than or equal to the segmentation threshold value as 0;
acquiring a binarization image corresponding to each iteration according to the segmentation threshold value after each iteration;
Acquiring the number of pixel points with pixel values of 1 in a binarized image as a first number;
Acquiring a first number of differences corresponding to the previous iteration and adjacent iteration of each iteration as a first difference, and acquiring a corresponding difference of the segmentation threshold as a second difference;
normalizing the first difference to obtain an analysis result;
Stopping iteration when the analysis result is smaller than a preset analysis threshold value and the second difference is smaller than or equal to a preset difference threshold value, and taking the segmentation threshold value of the current iteration as an optimal segmentation threshold value;
acquiring a binarization image corresponding to the optimal segmentation threshold value as a target binarization image;
Acquiring the number of all pixel points in the target binarized image as a second number;
Acquiring the ratio of the first quantity to the second quantity in the target binarized image as an evaluation value;
when the evaluation value is smaller than or equal to a preset evaluation value threshold value, the appearance quality of the power adapter is normal;
when the evaluation value is greater than a preset evaluation value threshold, the appearance quality of the power adapter is poor.
The invention has the following beneficial effects:
screening out target pixel points according to the gray value and the gradient amplitude of each pixel point in the gray image, and improving the efficiency of acquiring the initial segmentation threshold; according to the gradient and gray value of each target pixel point, obtaining a target feature vector of each target pixel point, and accurately distinguishing the scratch defect pixel point from the texture pixel point of the power adapter; further constructing a preset window for each target pixel point, acquiring similar characteristic values of each target pixel point and corresponding neighborhood target pixel points according to target characteristic vectors and gray values of the target pixel points and the neighborhood target pixel points in each preset window, and primarily judging the possibility that each target pixel point is a scratch defect pixel point; in order to obtain the initial segmentation threshold more accurately, further obtaining a target region according to the position distribution of the target pixel points; obtaining a possible defect value of each target pixel point according to the gray value, gradient and similar characteristic value of each target pixel point, the length of a target interval where each target pixel point is positioned and the width of a gray image, and accurately judging a scratch defect pixel point and a texture pixel point; according to the initial segmentation threshold, the probability of false detection and missed detection is reduced, so that the optimal segmentation threshold is accurately obtained through iterative threshold segmentation according to the initial segmentation threshold, the problem that the selection of the optimal segmentation threshold is inaccurate due to the fact that the pixel value difference between the scratch defect pixel point and the texture pixel point of the power adapter is too small is effectively solved, the phenomenon of over-segmentation and under-segmentation of the scratch defect is avoided, meanwhile, the accuracy of the segmentation of the scratch defect is improved, and the appearance quality of the power adapter is accurately detected.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a method for detecting appearance quality of a power adapter based on multi-feature fusion according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of the specific implementation, structure, characteristics and effects of the method for detecting the appearance quality of the power adapter based on multi-characteristic fusion according to the invention with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the method for detecting the appearance quality of the power adapter based on multi-feature fusion.
Referring to fig. 1, a flow chart of a method for detecting appearance quality of a power adapter based on multi-feature fusion according to an embodiment of the invention is shown, the method includes the following steps:
Step S1: a grayscale image of the power adapter is acquired.
Specifically, since the CCD appearance detector has high resolution and accuracy, the embodiment of the invention adopts the CCD appearance detector to acquire the appearance image of the power adapter, and performs denoising and graying treatment on the acquired appearance image of the power adapter to acquire the gray image of the power adapter. In order to detect the quality of the appearance of the power adapter more efficiently and accurately, the embodiment of the invention processes the gray image through a semantic segmentation algorithm to obtain the gray image only comprising the power adapter. The grayscale images appearing later all refer to grayscale images including only the power adapter. The denoising and graying processing and the semantic segmentation algorithm are all in the prior art, and are not described in detail.
The scene of the embodiment of the invention is as follows: the appearance of the black power adapter with textures is subjected to quality detection, wherein each texture is arranged in the same way and is parallel to and equal to one side of the power adapter. For better clarity of description, the embodiment of the invention uses one surface of the appearance of the power adapter as an example, and describes a method for detecting the quality of the appearance of the power adapter, wherein each surface of the power adapter is rectangular, that is, the gray image is rectangular. The scratch defect and texture on the appearance of the power adapter are brighter areas in the gray level image, namely the gray level value of the scratch defect pixel point and the texture pixel point in the gray level image is larger.
The aim of the embodiment of the invention is as follows: the gray value and gradient of the pixel points in the gray image are analyzed, and an initial segmentation threshold value of the gray image is obtained according to the characteristics corresponding to the scratch defects, so that the problem that the optimal segmentation threshold value obtained through iterative threshold segmentation is inaccurate and the segmentation of the scratch defects is inaccurate due to the fact that the difference of the gray value of the texture in the gray image and the gray value of the pixel points corresponding to the scratch defects is too small is solved. The accuracy of dividing the scratch defects in the gray level images is effectively improved, and then the appearance quality of the power adapter is accurately detected.
Step S2: screening out target pixel points according to the gray value and the gradient amplitude of each pixel point in the gray image; acquiring a target feature vector of each target pixel point according to the gradient and the gray value of each target pixel point; and constructing a preset window for each target pixel point, and acquiring similar characteristic values of each target pixel point and the corresponding neighborhood target pixel point according to the target characteristic vector and the gray value of the target pixel point and the neighborhood target pixel point in each preset window.
Specifically, as the shell of the power adapter is black, the scratch defect area and the texture area have obvious concave-convex areas, so that light rays can be diffusely reflected in the scratch defect area and the texture area, the scratch defect area and the texture area are brighter, the pixel value of a corresponding pixel point on a gray image is obviously larger, and the gradient amplitude is obviously changed. However, there is also a slight difference in the gray level of the scratch defective pixel point and the texture pixel point. The scratch defect is a long and narrow line on the power adapter, the diffuse reflection degree of light is high, and the brightness is higher than that of the texture of the power adapter, so that the gray value of the pixel point of the scratch defect in the gray image is slightly larger than that of the pixel point of the texture. The gray value of the pixel point in the region where the texture and scratch defect does not exist in the gray image is relatively small and the gradient amplitude hardly changes. Therefore, the pixel points corresponding to the scratch defects and the textures in the gray level image, namely the target pixel points, are obtained according to the gray level values and the gradient amplitude values of the pixel points, and further the target pixel points are analyzed, so that the possible defect value of each target pixel point is accurately obtained.
In the embodiment of the invention, the preset first gray value threshold is set to 30, and the preset first gradient amplitude threshold is set to 20, so that an operator can set the preset first gray value threshold and the preset first gradient amplitude threshold according to actual conditions, and the size of the preset first gray value threshold and the preset first gradient amplitude threshold is not limited. Taking an h pixel point in the gray level image as an example, when the gray level value of the h pixel point is greater than or equal to a preset first gray level value threshold value and the gradient amplitude value is greater than or equal to a preset first gradient amplitude value threshold value, the h pixel point is a target pixel point. And acquiring all target pixel points in the gray image according to the gray value and the gradient amplitude of each pixel point in the gray image. The method for acquiring the gradient amplitude is the prior art, and is not described in detail.
In order to better distinguish scratch defect pixel points and texture pixel points, the embodiment of the invention firstly obtains the target feature vector of each target pixel point, then constructs a preset window for each target pixel point, obtains the similar feature value of each target pixel point and the corresponding neighborhood target pixel point according to the target feature vector and the gray value of the target pixel point and the neighborhood target pixel point in each preset window, and preliminarily judges the possibility that each target pixel point is the scratch defect pixel point according to the similar feature value. The specific method for acquiring the similar characteristic values is as follows:
(1) And obtaining the target feature vector.
For the texture pixel points, the gradient direction of each texture pixel point is necessarily vertical to the direction in which the texture is located. The direction of each stripe distribution can be set by the practitioner according to the actual situation, and is not limited herein. For the scratch defect pixel points, the gradient direction of each scratch defect pixel point is random, so that the embodiment of the invention obtains the included angle between the gradient direction and the vertical direction of each target pixel point as a first included angle. The larger the first included angle is, the more likely the corresponding target pixel point is the scratch defect pixel point. And acquiring the gray value, gradient amplitude and first included angle of each target pixel point, constructing a target feature vector of each target pixel point, and preparing for acquiring the possible defect value of each target pixel point.
Preferably, the method for acquiring the target feature vector comprises the following steps: for any target pixel point, acquiring an included angle between the gradient direction of the target pixel point and a preset direction as a first included angle of the target pixel point; the result of normalizing the gray value of the target pixel point is taken as a first result; the result of normalizing the gradient amplitude of the target pixel point is taken as a second result; the result of normalizing the first included angle of the target pixel point is used as a third result; taking the product of the first preset weight and the first result as a first value; taking the product of the second preset weight and the second result as a second value; taking the product of the third preset weight and the third result as a third value; and the first value represents the x value in the three-dimensional coordinate system, the second value represents the y value in the three-dimensional coordinate system, the third value represents the z value in the three-dimensional coordinate system, and the target feature vector of the target pixel point in the three-dimensional coordinate system is obtained.
Taking the ith target pixel point as an example, an included angle between the gradient direction of the ith target pixel point and the vertical direction, i.e. the preset direction, is obtained, i.e. the first included angle of the ith target pixel pointWherein the first included angle/>The range of the value is set to be 0 to 360 degrees according to the anticlockwise direction, and the operator can set the size and the direction of the first included angle according to the actual situation, which is not limited herein. Obtaining the gray value/>, of the ith target pixel pointAnd gradient magnitude/>The gray value/>, of the ith target pixel pointGradient amplitude/>And a first included angle/>Respectively carrying out normalization processing, and respectively obtaining the following results: first result/>Second result/>And third result/>. In the embodiment of the invention, the first preset weight is set to 0.4, the second preset weight is set to 0.4, and the third preset weight is set to 0.2, so that the magnitude of the first preset weight, the second preset weight and the third preset weight can be set according to actual conditions by an implementer, and the implementation is not limited herein. Obtaining a first value, i.e. the product of the first preset weight and the first result, as/>The product of the second preset weight and the second result is a second value of/>The product of the third preset weight and the third result is a third value of/>. Will/>、/>And/>Sequentially representing the x value, the y value and the z value in the three-dimensional coordinate system to further obtain the target feature vector/>, of the ith target pixel point, in the three-dimensional coordinate system。
And obtaining the target feature vector of each target pixel point according to the method for obtaining the target feature vector of the ith target pixel point.
(2) And obtaining similar characteristic values.
Each texture in the gray image is a horizontal line segment, and the target feature vector and the gray value of each texture pixel point are similar. The shape of the scratch defect in the gray image is not fixed, and the target feature vector of the scratch defect pixel point is different. Therefore, in the embodiment of the present invention, a preset window is built with each target pixel point as a center, and the size of the preset window is set as followsThe size of the preset window can be set by the practitioner according to the actual situation, and the method is not limited herein. Obtaining the similarity between the central target pixel point of each preset window and the target feature vector of each neighborhood target pixel point through cosine similarity; and obtaining similar characteristic values of the central target pixel point and the corresponding neighborhood target pixel points according to the similarity and the gray value difference of the central target pixel point of each preset window and each neighborhood target pixel point. The cosine similarity is the prior art, and will not be described in detail.
Taking the ith target pixel point as an example, taking the ith target pixel point as a center, constructing a preset window of the ith target pixel point, and taking target pixel points which are not the ith target pixel point in the window as neighborhood target pixel points of the ith target pixel point. If the ith target pixel point is a boundary pixel point of the gray level image, only analyzing a preset window part in the gray level image. And obtaining the similarity between the ith target pixel point and the corresponding target feature vector of each neighborhood target pixel point through cosine similarity. And further, according to the similarity and the absolute value of the difference value of the gray values of the ith target pixel point and each corresponding neighborhood target pixel point, acquiring a calculation formula of the similarity characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point as follows:
In the method, in the process of the invention, The similar characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point; /(I)The number of the neighbor target pixel points in the preset window of the ith target pixel point is set; /(I)The similarity between the ith target pixel point and the target feature vector of the (y) th neighborhood target pixel point in a preset window of the ith target pixel point is obtained; /(I)The gray value of the ith target pixel point; /(I)The gray value of the target pixel point in the y neighborhood in the preset window of the i target pixel point is obtained; /(I)Is a preset constant, and is larger than 0; /(I)As a function of absolute value.
Embodiments of the invention willSet to 1, avoid denominator to 0, and the practitioner can set/>, according to the actual situationIs not limited herein.
It should be noted that the number of the substrates,The larger the target feature vector of the ith target pixel point and the corresponding y-th neighborhood target pixel point is, the more likely the ith target pixel point and the corresponding y-th neighborhood target pixel point are texture pixel points,/>, the more the target feature vector of the ith target pixel point and the corresponding y-th neighborhood target pixel point is the sameThe larger; /(I)The smaller the i-th target pixel point and the corresponding y-th neighborhood target pixel point are, the more likely the i-th target pixel point and the corresponding y-th neighborhood target pixel point are texture pixel points,/>The larger; thus,/>The larger the i-th target pixel is, the less likely it is to be a scratch defective pixel.
And obtaining the similar characteristic value of each target pixel point and the corresponding neighborhood target pixel point according to the method for obtaining the similar characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point.
Step S3: acquiring at least one target area according to the position distribution of the target pixel points; and obtaining the possible defect value of each target pixel point according to the gray value, gradient and the similar characteristic value of each target pixel point, the length of the target interval where each target pixel point is positioned and the width of the gray image.
Specifically, the texture area and the scratch defect area are areas formed by the target pixel points, and in actual situations, the situation that part of the texture pixel points and the scratch defect pixel points cannot be identified possibly exists, so that the texture area or the scratch defect area is interrupted, and the distinction of the texture pixel points and the scratch defect pixel points is affected. Therefore, in the embodiment of the invention, each target pixel point is used as a seed point, and the target region is acquired through a region growing algorithm. The region growing algorithm is in the prior art, and will not be described in detail. The stopping condition of the area growth algorithm in the embodiment of the invention is as follows: the absolute value of the difference value of the gray values between two adjacent target pixel points is larger than the preset second gray value threshold value, the preset second gray value threshold value is set to be 2, and an implementer can set the magnitude of the preset second gray value threshold value according to actual conditions, and the implementation is not limited. The acquired target area includes each texture area and each scratch defect area. Since the line segment where the length of each texture area is known to be parallel and equal to one side of the gray image and the distribution of each scratch defect area is uncertain, the embodiment of the invention obtains the circumscribed rectangle of each target area, takes the maximum side length of the circumscribed rectangle as the length of the corresponding target area and takes the side length corresponding to the texture in the gray image as the width of the gray image. The more the length of the target area is equal to the width of the gray level image, the more the corresponding target area is likely to be a texture area, and in practical situations, the length of the target area corresponding to a certain scratch defect area is equal to the width of the gray level image, at this time, the scratch defect area and the texture area cannot be distinguished only according to the length of the target area and the width of the gray level image, and therefore the scratch defect pixel point cannot be accurately judged. Therefore, the embodiment of the invention acquires the possible defect value of each target pixel point according to the gray value, gradient and similar characteristic value of each target pixel point, the length of the target interval where each target pixel point is positioned and the width of the gray image.
As an example, taking the ith target pixel point in step S2 as an example, the first included angle of the ith target pixel point is obtained according to the gradient direction of the ith target pixel point. According to the first included angle, the gray value and the similar characteristic value of the ith target pixel point, the length of the target interval where the ith target pixel point is positioned and the width of the gray image, the calculation formula for obtaining the possible defect value of the ith target pixel point is as follows:
In the method, in the process of the invention, The possible defect value of the ith target pixel point; /(I)The first included angle is the first included angle of the ith target pixel point; The gray value of the ith target pixel point; /(I) The length of the target interval where the ith target pixel point is located; d is the width of the gray scale image; /(I)The similar characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point; exp is an exponential function based on a natural constant; norm is a normalization function; /(I)As a function of absolute value.
It should be noted that the number of the substrates,The larger the i-th target pixel point is, the more probable the i-th target pixel point is a scratch defect pixel point,/>The larger; /(I)The larger the i-th target pixel is, the more likely it is a scratch defect pixel,/>The larger; /(I)The larger the description/>The more different from D, the more likely the target area where the ith target pixel is located is a scratch defect area, the more likely the ith target pixel is a scratch defect pixel,/>The larger the/>The larger; /(I)The smaller the i-th target pixel point is, the more probable the i-th target pixel point is a scratch defect pixel point,/>The larger; thus,/>The larger the i-th target pixel is, the more likely the scratch-defective pixel is.
And obtaining the possible defect value of each target pixel point according to the method for obtaining the possible defect value of the ith target pixel point.
Step S4: acquiring an initial segmentation threshold according to the possible defect value and the gray value of each target pixel point; and according to the initial segmentation threshold, obtaining an optimal segmentation threshold through iterative threshold segmentation, and detecting the quality of the appearance of the power adapter.
Specifically, by analyzing the possible defect value of each target pixel point, the initial segmentation threshold can be acquired more accurately, the initial segmentation threshold can be effectively prevented from falling into a local minimum, the probability of false detection and missing detection is reduced, the accuracy of iterative threshold segmentation is effectively improved, the optimal segmentation threshold is further acquired accurately, and the appearance of the power adapter is detected accurately. The iterative threshold segmentation is in the prior art, and is not described in detail.
Preferably, the method for acquiring the initial segmentation threshold value is as follows: acquiring the average value of possible defect values of each target pixel point to be used as a defect distinguishing threshold; when the possible defect value is larger than the defect distinguishing threshold value, the corresponding target pixel point is used as a first class pixel point; when the possible defect value is smaller than or equal to the defect distinguishing threshold value, the corresponding target pixel point is used as a second-class pixel point; the first class pixel substance may default to scratch-defective pixel and the second class pixel substance may default to texture pixel. And further, gray values and possible defect values in the first class pixel points and the second class pixel points are analyzed, so that the obtained initial segmentation threshold value is more accurate. Acquiring a gray value average value and a defect possible value average value of first class pixel points, and respectively serving as a first gray characteristic value and a first defect characteristic value; taking the product of the first gray characteristic value and the first defect characteristic value as a first characteristic value; acquiring a gray value average value and a defect possible value average value of the second class of pixel points to respectively serve as a second gray characteristic value and a second defect characteristic value; taking the product of the second gray characteristic value and the second defect characteristic value as a second characteristic value; and taking the addition result of the first characteristic value and the second characteristic value as an initial segmentation threshold value. The calculation formula for acquiring the initial segmentation threshold value is as follows:
In the method, in the process of the invention, Is an initial segmentation threshold; /(I)Is a first defect characteristic value; /(I)Is a first gray scale feature value; /(I)Is a second defect characteristic value; /(I)The second gray characteristic value; /(I)Is a first characteristic value; /(I)Is the second eigenvalue.
The first characteristic valueThe larger the/>The larger; second eigenvalue/>The larger the size of the container,The larger; /(I)The larger the gray value of the pixel point indirectly indicating the scratch defect is larger.
And/>As weights, respectively pair/>And/>Correction is performed such that/>More accurate. And then according to the initial segmentation threshold/>Obtaining a segmentation threshold value after each iteration through iterative threshold segmentation; setting the pixel value of the pixel point with the gray value larger than the segmentation threshold value in the gray image as 1, and setting the pixel value of the pixel point with the gray value smaller than or equal to the segmentation threshold value as 0; acquiring a binarization image corresponding to each iteration according to the segmentation threshold value after each iteration; acquiring the number of pixel points with pixel values of 1 in a binarized image as a first number; obtaining the absolute value of the difference value of the first quantity corresponding to the previous iteration and adjacent iteration of each iteration as a first difference, and the absolute value of the difference value of the corresponding segmentation threshold value as a second difference; and normalizing the first difference to obtain an analysis result. In the embodiment of the invention, the preset analysis threshold value is set to be 0.005 and the preset difference threshold value is set to be 0.5, and an implementer can set the preset analysis threshold value and the preset difference threshold value according to actual conditions without limitation. And stopping iteration when the analysis result is smaller than a preset analysis threshold and the second difference is smaller than or equal to a preset difference threshold, and taking the segmentation threshold of the current iteration as the optimal segmentation threshold. The formula for obtaining the analysis result in the embodiment of the invention is as follows: /(I)In the/>For the analysis result corresponding to the kth iteration,/>For the first number corresponding to the kth iteration,/>For the first number corresponding to the k-1 th iteration,/>As a function of absolute value; /(I)Is the first difference. First difference/>The larger the difference between the segmentation threshold values corresponding to the kth iteration and the kth-1 iteration is, the less likely the segmentation threshold value corresponding to the kth iteration is the optimal segmentation threshold value; first difference/>The smaller the difference between the segmentation threshold values for the kth iteration and the kth-1 iteration is, the more likely the segmentation threshold value for the kth iteration is the optimal segmentation threshold value. So far, the optimal segmentation threshold is accurately acquired.
Acquiring a binarization image corresponding to the optimal segmentation threshold value as a target binarization image; acquiring the number of all pixel points in the target binarized image as a second number; and acquiring the ratio of the first quantity to the second quantity in the target binarized image as an evaluation value. In the embodiment of the invention, the preset evaluation value threshold is set to be 0.02, and the magnitude of the preset evaluation value threshold can be set by an implementer according to actual conditions, so that the method is not limited. When the evaluation value is smaller than or equal to a preset evaluation value threshold, the appearance quality of the power adapter is normal, and the power adapter can be directly put into the market for use; when the evaluation value is larger than a preset evaluation value threshold, the appearance quality of the power adapter is poor, and the corresponding power adapter shell needs to be reprocessed, so that scratch defects affecting normal use are eliminated.
The present invention has been completed.
In summary, the embodiment of the invention acquires the gray image of the power adapter; screening out target pixel points according to the gray value and the gradient amplitude value, and obtaining target feature vectors of the target pixel points; obtaining a similar characteristic value according to the target characteristic vector and the gray value of the target pixel point; and obtaining possible defect values according to the gray value, gradient and similar characteristic values of the target pixel point, the length of a target interval where the target pixel point is positioned and the width of a gray image, further obtaining an initial segmentation threshold value, determining an optimal segmentation threshold value through iterative threshold segmentation, and detecting the quality of the appearance of the power adapter. According to the invention, the initial segmentation threshold value is accurately obtained, and then the optimal segmentation threshold value is accurately determined to accurately segment the scratch defect in the gray level image through iterative threshold segmentation, so that the quality of the appearance of the power adapter is accurately detected.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (10)
1. The method for detecting the appearance quality of the power adapter based on multi-feature fusion is characterized by comprising the following steps of:
Acquiring a gray image of a power adapter;
Screening out target pixel points according to the gray value and the gradient amplitude of each pixel point in the gray image; acquiring a target feature vector of each target pixel point according to the gradient and the gray value of each target pixel point; constructing a preset window for each target pixel point, and acquiring a similar characteristic value of each target pixel point and a corresponding neighborhood target pixel point according to a target characteristic vector and a gray value of the target pixel point and the neighborhood target pixel point in each preset window;
Acquiring at least one target area according to the position distribution of the target pixel points; obtaining a possible defect value of each target pixel point according to the gray value, gradient and the similar characteristic value of each target pixel point, the length of a target area where each target pixel point is positioned and the width of a gray image;
acquiring an initial segmentation threshold according to the possible defect value and the gray value of each target pixel point; and according to the initial segmentation threshold, acquiring an optimal segmentation threshold through an iterative threshold segmentation algorithm, and detecting the quality of the appearance of the power adapter.
2. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 1, wherein the method for obtaining the target pixel point is as follows:
For any pixel point, when the gray value of the pixel point is greater than or equal to a preset first gray value threshold value and the gradient amplitude is greater than or equal to a preset first gradient amplitude threshold value, the pixel point is a target pixel point.
3. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 1, wherein the method for obtaining the target feature vector of each target pixel point according to the gradient and the gray value of each target pixel point is as follows:
For any target pixel point, acquiring an included angle between the gradient direction of the target pixel point and a preset direction as a first included angle of the target pixel point;
Taking the result of normalizing the gray value of the target pixel point as a first result;
the result of normalizing the gradient amplitude of the target pixel point is taken as a second result;
The result of normalizing the first included angle of the target pixel point is taken as a third result;
Taking the product of the first preset weight and the first result as a first value;
taking the product of the second preset weight and the second result as a second value;
taking the product of the third preset weight and the third result as a third value;
And the first value represents the x value in the three-dimensional coordinate system, the second value represents the y value in the three-dimensional coordinate system, the third value represents the z value in the three-dimensional coordinate system, and the target feature vector of the target pixel point in the three-dimensional coordinate system is obtained.
4. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 1, wherein the method for constructing a preset window for each target pixel point and obtaining the similar feature value of each target pixel point and the corresponding neighborhood target pixel point according to the target feature vector and the gray value of the target pixel point and the neighborhood target pixel point in each preset window is as follows:
Constructing a preset window by taking each target pixel point as a center, and acquiring the similarity between the central target pixel point of each preset window and the target feature vector of each neighborhood target pixel point through cosine similarity;
And obtaining similar characteristic values of the central target pixel point and the corresponding neighborhood target pixel points according to the similarity and the gray value difference of the central target pixel point of each preset window and each neighborhood target pixel point.
5. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 4, wherein the calculation formula of the similar feature value is as follows:
In the method, in the process of the invention, The similar characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point; /(I)The number of the neighbor target pixel points in the preset window of the ith target pixel point is set; /(I)The similarity between the ith target pixel point and the target feature vector of the (y) th neighborhood target pixel point in a preset window of the ith target pixel point is obtained; /(I)The gray value of the ith target pixel point; /(I)The gray value of the target pixel point in the y neighborhood in the preset window of the i target pixel point is obtained; /(I)Is a preset constant, and is larger than 0; /(I)As a function of absolute value.
6. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 1, wherein the method for acquiring the target area is as follows:
Taking each target pixel point as a seed point, and acquiring a target region through a region growing algorithm; wherein, the stopping condition of the region growing algorithm is as follows: the gray value difference between two adjacent target pixel points is larger than a preset second gray value threshold value.
7. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 3, wherein the calculation formula of the possible defect value is as follows:
In the method, in the process of the invention, The possible defect value of the ith target pixel point; /(I)The first included angle is the first included angle of the ith target pixel point; /(I)The gray value of the ith target pixel point; /(I)The length of the target area where the ith target pixel point is located; d is the width of the gray scale image; the similar characteristic value of the ith target pixel point and the corresponding neighborhood target pixel point; exp is an exponential function based on a natural constant; norm is a normalization function; /(I) As a function of absolute value.
8. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 1, wherein the method for obtaining the length of the target area is as follows: for any target area, acquiring a circumscribed rectangle of the target area, determining the side length of the circumscribed rectangle, and taking the maximum side length as the length of the target area;
The width of the gray image refers to the side length corresponding to the texture in the gray image.
9. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 1, wherein the method for obtaining the initial segmentation threshold is as follows:
Acquiring an average value of possible defect values of a plurality of target pixel points as a defect distinguishing threshold;
when the possible defect value is larger than the defect distinguishing threshold value, the corresponding target pixel point is used as a first class pixel point;
when the possible defect value is smaller than or equal to the defect distinguishing threshold value, the corresponding target pixel point is used as a second-class pixel point;
Acquiring a gray value average value and a defect possible value average value of first class pixel points, and respectively serving as a first gray characteristic value and a first defect characteristic value;
taking the product of the first gray characteristic value and the first defect characteristic value as a first characteristic value;
Acquiring a gray value average value and a defect possible value average value of the second class of pixel points to respectively serve as a second gray characteristic value and a second defect characteristic value;
Taking the product of the second gray characteristic value and the second defect characteristic value as a second characteristic value;
and taking the addition result of the first characteristic value and the second characteristic value as an initial segmentation threshold value.
10. The method for detecting the appearance quality of the power adapter based on multi-feature fusion according to claim 1, wherein the method for detecting the appearance quality of the power adapter by obtaining an optimal segmentation threshold through an iterative threshold segmentation algorithm according to an initial segmentation threshold is as follows:
Based on the initial segmentation threshold, obtaining a segmentation threshold after each iteration through an iterative threshold segmentation algorithm;
Setting the pixel value of the pixel point with the gray value larger than the segmentation threshold value in the gray image as 1, and setting the pixel value of the pixel point with the gray value smaller than or equal to the segmentation threshold value as 0;
acquiring a binarization image corresponding to each iteration according to the segmentation threshold value after each iteration;
Acquiring the number of pixel points with pixel values of 1 in a binarized image as a first number;
Acquiring a first number of differences corresponding to the previous iteration and adjacent iteration of each iteration as a first difference, and acquiring a corresponding difference of the segmentation threshold as a second difference;
Taking the result of normalizing the first difference as an analysis result;
Stopping iteration when the analysis result is smaller than a preset analysis threshold value and the second difference is smaller than or equal to a preset difference threshold value, and taking the segmentation threshold value of the current iteration as an optimal segmentation threshold value;
acquiring a binarization image corresponding to the optimal segmentation threshold value as a target binarization image;
Acquiring the number of all pixel points in the target binarized image as a second number;
Acquiring the ratio of the first quantity to the second quantity in the target binarized image as an evaluation value;
when the evaluation value is smaller than or equal to a preset evaluation value threshold value, the appearance quality of the power adapter is normal;
when the evaluation value is greater than a preset evaluation value threshold, the appearance quality of the power adapter is poor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311687400.XA CN117408995B (en) | 2023-12-11 | 2023-12-11 | Power adapter appearance quality detection method based on multi-feature fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311687400.XA CN117408995B (en) | 2023-12-11 | 2023-12-11 | Power adapter appearance quality detection method based on multi-feature fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117408995A CN117408995A (en) | 2024-01-16 |
| CN117408995B true CN117408995B (en) | 2024-05-24 |
Family
ID=89494683
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311687400.XA Active CN117408995B (en) | 2023-12-11 | 2023-12-11 | Power adapter appearance quality detection method based on multi-feature fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117408995B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117745724B (en) * | 2024-02-20 | 2024-04-26 | 高唐县瑞景精密机械有限公司 | Stone polishing processing defect region segmentation method based on visual analysis |
| CN117764992B (en) * | 2024-02-22 | 2024-04-30 | 山东乔泰管业科技有限公司 | Plastic pipe quality detection method based on image processing |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2016109437A (en) * | 2014-12-02 | 2016-06-20 | 株式会社リコー | Image processing device and method, and defect inspection device |
| CN114627111A (en) * | 2022-05-12 | 2022-06-14 | 南通英伦家纺有限公司 | Textile defect detection and identification device |
| CN114913365A (en) * | 2022-04-22 | 2022-08-16 | 海门王巢家具制造有限公司 | Artificial board quality classification method and system based on machine vision |
| CN115619793A (en) * | 2022-12-21 | 2023-01-17 | 深圳市澳博森科技有限公司 | Power adapter appearance quality detection method based on computer vision |
| CN116703909A (en) * | 2023-08-07 | 2023-09-05 | 威海海泰电子有限公司 | Intelligent detection method for production quality of power adapter |
| CN117173184A (en) * | 2023-11-03 | 2023-12-05 | 济宁市市政园林养护中心 | Road construction quality detection method and system based on artificial intelligence |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SG10201405182WA (en) * | 2014-08-25 | 2016-03-30 | Univ Singapore Technology & Design | Method and system |
| WO2016207875A1 (en) * | 2015-06-22 | 2016-12-29 | Photomyne Ltd. | System and method for detecting objects in an image |
-
2023
- 2023-12-11 CN CN202311687400.XA patent/CN117408995B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2016109437A (en) * | 2014-12-02 | 2016-06-20 | 株式会社リコー | Image processing device and method, and defect inspection device |
| CN114913365A (en) * | 2022-04-22 | 2022-08-16 | 海门王巢家具制造有限公司 | Artificial board quality classification method and system based on machine vision |
| CN114627111A (en) * | 2022-05-12 | 2022-06-14 | 南通英伦家纺有限公司 | Textile defect detection and identification device |
| CN115619793A (en) * | 2022-12-21 | 2023-01-17 | 深圳市澳博森科技有限公司 | Power adapter appearance quality detection method based on computer vision |
| CN116703909A (en) * | 2023-08-07 | 2023-09-05 | 威海海泰电子有限公司 | Intelligent detection method for production quality of power adapter |
| CN117173184A (en) * | 2023-11-03 | 2023-12-05 | 济宁市市政园林养护中心 | Road construction quality detection method and system based on artificial intelligence |
Non-Patent Citations (1)
| Title |
|---|
| 基于梯度阈值对象提取的图像检索技术研究;周艺华 等;《计算机工程与设计》;20130731;第178-182页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117408995A (en) | 2024-01-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117408995B (en) | Power adapter appearance quality detection method based on multi-feature fusion | |
| CN115375676B (en) | Stainless steel product quality detection method based on image recognition | |
| Liu et al. | A detection and recognition system of pointer meters in substations based on computer vision | |
| CN108629775B (en) | Thermal state high-speed wire rod surface image processing method | |
| US5081689A (en) | Apparatus and method for extracting edges and lines | |
| CN113570605B (en) | Defect detection method and system based on liquid crystal display panel | |
| CN116843678B (en) | Hard carbon electrode production quality detection method | |
| CN115020267B (en) | Semiconductor surface defect detection method | |
| CN112651968B (en) | Wood board deformation and pit detection method based on depth information | |
| CN115170570B (en) | Fabric fuzzing and pilling detection method based on gray level run matrix | |
| CN117115153B (en) | Intelligent printed circuit board quality detection method based on visual assistance | |
| Said et al. | Robust automatic void detection in solder balls | |
| CN115272256A (en) | Sub-pixel level sensing optical fiber path Gaussian extraction method and system | |
| CN116823822B (en) | Ship sheet metal part welding defect detection method based on visual characteristics | |
| CN108335310B (en) | Portable grain shape and granularity detection method and system | |
| CN115953409A (en) | Injection molding surface defect detection method based on image processing | |
| CN115330802A (en) | Carbon fiber composite material gas cylinder X-ray image debonding defect extraction method | |
| CN115471476A (en) | Method, device, equipment and medium for detecting component defects | |
| CN116612112B (en) | Visual inspection method for surface defects of bucket | |
| CN113012127A (en) | Cardiothoracic ratio measuring method based on chest medical image | |
| CN113554688B (en) | O-shaped sealing ring size measurement method based on monocular vision | |
| CN113822361B (en) | SAR image similarity measurement method and system based on Hamming distance | |
| JP5452035B2 (en) | Defect inspection method and defect inspection apparatus | |
| CN116758086B (en) | Bulldozer part quality detection method based on image data | |
| CN117830312B (en) | Alloy crack nondestructive testing method based on machine vision |
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 |