CN117649412A - Aluminum material surface quality detection method - Google Patents
Aluminum material surface quality detection method Download PDFInfo
- Publication number
- CN117649412A CN117649412A CN202410121803.6A CN202410121803A CN117649412A CN 117649412 A CN117649412 A CN 117649412A CN 202410121803 A CN202410121803 A CN 202410121803A CN 117649412 A CN117649412 A CN 117649412A
- Authority
- CN
- China
- Prior art keywords
- gray
- region
- point
- defect
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- XAGFODPZIPBFFR-UHFFFAOYSA-N aluminium Chemical compound [Al] XAGFODPZIPBFFR-UHFFFAOYSA-N 0.000 title claims abstract description 57
- 229910052782 aluminium Inorganic materials 0.000 title claims abstract description 56
- 239000000463 material Substances 0.000 title claims abstract description 39
- 238000001514 detection method Methods 0.000 title description 11
- 230000007547 defect Effects 0.000 claims abstract description 103
- 230000002159 abnormal effect Effects 0.000 claims abstract description 91
- 230000005856 abnormality Effects 0.000 claims abstract description 41
- 238000000034 method Methods 0.000 claims abstract description 35
- 230000035772 mutation Effects 0.000 claims abstract description 18
- 238000001914 filtration Methods 0.000 claims abstract description 5
- 238000007781 pre-processing Methods 0.000 claims abstract description 5
- 230000011218 segmentation Effects 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000012163 sequencing technique Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 4
- 229910052751 metal Inorganic materials 0.000 description 4
- 239000002184 metal Substances 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000031068 symbiosis, encompassing mutualism through parasitism Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 150000002739 metals Chemical class 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Landscapes
- Image Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image data processing, in particular to a method for detecting the surface quality of aluminum materials. The method comprises the following steps: collecting an image of the surface of the aluminum material and preprocessing to obtain a surface gray level image of the aluminum material; obtaining an abnormal region corresponding to a gray maximum point according to the edges of the regions corresponding to the first mutation point and the second mutation point in the gray variance sequence; obtaining gradient abnormality of the abnormal region according to the average gradient of the edge of the abnormal region; the product of the number of elements in the direction sequence and the point-to-gray level abnormality is the gray level symbiotic abnormality of the abnormal region; the defect area with the defect credibility greater than the lower quartile is a suspected dirty point defect area; filtering other areas except for the suspected dirty point defect area in the surface gray scale map to obtain a filtered image; and detecting the quality of the surface of the aluminum product according to the filtered image. The invention can accurately detect the surface quality of the aluminum material.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a method for detecting the surface quality of aluminum materials.
Background
In the industrial production process of metals, the surfaces are easily defective due to the influence of various factors, and the quality and the appearance are seriously influenced. Quality inspection is a crucial step in order to guarantee the market and public praise of the product. Aluminum is an important metal, and has dirty point defects. With the development of machine vision, the prior art detects dirty points through threshold segmentation, but because the degree of distinction between a surface dirty point defect area and a non-defect area is not high, noise with little difference from the gray level of the dirty points exists on the surface, the detection accuracy is low, and the technology needs to be optimized.
Dirty points on the surface of the aluminum product have various expression forms, and the dirty points caused by breakage are not obvious and are easy to be confused with noise in a normal surface area, so that the problem of inaccurate detection of the surface quality of the aluminum product is caused.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a method for detecting the surface quality of an aluminum material, which adopts the following technical scheme:
the embodiment of the invention provides a method for detecting the surface quality of an aluminum material, which comprises the following steps:
collecting an image of the surface of the aluminum material and preprocessing to obtain a surface gray level image of the aluminum material; obtaining a suspected defect gray scale map in the surface gray scale map; obtaining at least two gray maximum points in the suspected defect gray map; performing region expansion at least twice by taking the gray maximum point as a center to obtain a region sequence corresponding to the gray maximum point; calculating the gray variance of each region of the region sequence corresponding to the gray maximum point to form a gray variance sequence; obtaining an abnormal region corresponding to a gray maximum point according to the edges of the regions corresponding to the first mutation point and the second mutation point in the gray variance sequence;
obtaining gradient abnormality of the abnormal region according to the average gradient of the edge of the abnormal region; setting a threshold value, obtaining gray level co-occurrence matrixes of each abnormal region in four directions, and obtaining point-to-gray level abnormality of the abnormal region based on point pairs with absolute values of gray level differences of point pairs in the gray level co-occurrence matrixes in each direction being larger than the threshold value; obtaining a direction sequence based on the direction of the connecting line of the pixel points in the point pair with the absolute value of the gray difference of the point pair larger than the threshold value in the gray level co-occurrence matrix in each direction; the product of the number of elements in the direction sequence and the point-to-gray level abnormality is the gray level symbiotic abnormality of the abnormal region;
obtaining defect credibility according to gradient abnormality and gray level symbiotic abnormality of an abnormal region; screening the abnormal region according to the defect reliability to obtain a suspected dirty point defect region; filtering other areas except for the suspected dirty point defect area in the surface gray scale map to obtain a filtered image; and detecting the quality of the surface of the aluminum product according to the filtered image.
Preferably, acquiring an image of the surface of the aluminum material and performing pretreatment to obtain a surface gray scale image of the aluminum material comprises: carrying out semantic segmentation on the image on the surface of the aluminum material, setting the pixel value of the background area to be 0, and keeping the pixel value of the aluminum material area unchanged to obtain the image after semantic segmentation; and carrying out graying treatment on the semantically segmented image to obtain a surface gray scale image.
Preferably, obtaining at least two gray maxima points in the gray map of suspected defects includes:
setting a sliding window with a preset size, and when the sliding window slides on the suspected defect gray scale map, obtaining a pixel point with the maximum gray scale value in each sliding window as a gray scale maximum value point.
Preferably, obtaining the abnormal region corresponding to the gray maximum point according to the edges of the regions corresponding to the first mutation point and the second mutation point in the gray variance sequence includes: the edge of the region in the region sequence corresponding to the first mutation point is the inner edge of the abnormal region, and the edge of the region in the region sequence corresponding to the second mutation point is the outer edge of the abnormal region.
Preferably, obtaining the gradient abnormality of the abnormal region from the average gradient of the edge of the abnormal region includes: the ratio of the average gradient value of the inner edge to the average gradient value of the outer edge of the abnormal region is the gradient abnormality of the abnormal region.
Preferably, the point-to-gray anomaly is:
wherein,point-to-gray scale anomalies representing the s-th anomaly region; />Respectively representing gray level co-occurrence matrixes in four directions of an abnormal region; />Representing a threshold value; />A set of point pairs representing a gray level co-occurrence matrix of the s-th anomaly region; />An absolute value representing a difference in gray values of two pixel points in a pair of points; />The frequency of occurrence of the absolute value of the difference in gray value of two pixel points in a pair is represented.
Preferably, the defect confidence level is:
wherein,representing the defect credibility of the s-th abnormal region; />Gradient abnormality representing the s-th abnormality region; />Gray level co-occurrence abnormality representing the s-th abnormal region; />Expressed in natural constante is the underlying exponential function.
Preferably, the quality detection of the aluminum surface based on the filtered image includes: and (3) re-carrying out threshold segmentation on the filtered image to obtain a highlight region in the dirty point defects in a damaged state, wherein the highlight region marked after segmentation is the region where the dirty point defects are located, and outputting the position parameters of the region where the defects are located.
Preferably, the method for acquiring the gray-scale map of the suspected defect comprises the following steps:
obtaining a gradient image of the surface gray level map, and dividing the gradient image to obtain a mask image; and multiplying the mask image by the surface gray scale map to obtain a suspected defect gray scale map.
Preferably, the method for acquiring the suspected dirty point defect region includes:
sequencing the defect credibility of all abnormal areas, and obtaining the lower quartile based on sequencing results; the abnormal region with the defect reliability degree larger than the lower quartile is a suspected dirty point defect region.
The embodiment of the invention has at least the following beneficial effects: the method is based on Sobel operator and Ojin segmentation to obtain the suspected defect gray map, and the complete abnormal region is obtained by region growth, so that a foundation is laid for further judging whether the defect is a dirty point defect, and the omission ratio of the defect is reduced. The defect reliability is obtained by calculating the abnormality of the abnormal region on the gradient and gray level symbiosis, the noise and the dirty points which are the same as the abnormal region are distinguished by the indexes, the interference of the noise on the dirty point defect detection is reduced, and the defect detection accuracy is improved.
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 the surface quality of an aluminum product 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 specific implementation, structure, characteristics and effects of an aluminum material surface quality detection method according to the invention with reference to the accompanying drawings and preferred embodiments. 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 surface quality of the aluminum material provided by the invention with reference to the accompanying drawings.
Examples:
the main application scene of the invention is as follows: in the quality inspection stage of the aluminum metal surface, the defects of unobvious dirty points caused by breakage are easily confused with noise points on the product surface, and the omission ratio and the false detection ratio are higher.
The main purpose of the invention is as follows: for the dirty point defects which are difficult to detect on the surface of the aluminum material, the abnormal areas with the defects are searched, the corresponding defect credibility is calculated, partial noise points are selectively removed, the dirty point areas which are wanted to be found are highlighted, the image enhancement is realized, and the defect omission ratio is reduced.
Referring to fig. 1, a method flowchart of a method for detecting surface quality of an aluminum material according to an embodiment of the invention is shown, and the method includes the following steps:
s1, collecting an image of the surface of an aluminum product and preprocessing to obtain a surface gray scale image of the aluminum product; obtaining a gradient image of the surface gray level map, and dividing the gradient image to obtain a mask image; multiplying the mask image with the surface gray level map to obtain a suspected defect gray level map; obtaining at least two gray maximum points in the suspected defect gray map; performing region expansion at least twice by taking the gray maximum point as a center to obtain a region sequence corresponding to the gray maximum point; calculating the gray variance of each region of the region sequence corresponding to the gray maximum point to form a gray variance sequence; and obtaining an abnormal region corresponding to the gray maximum value point according to the edges of the regions corresponding to the first mutation point and the second mutation point in the gray variance sequence.
The aluminum product after production is placed on a conveying line to wait for quality inspection, so a camera acquisition system is placed above the conveying line, and images of the surface of the aluminum product are acquired downwards. Because the acquired image contains background images such as a production line, the image is subjected to semantic segmentation, the pixel value of a background region which is not interested is set to be 0, and the pixel value of an aluminum region which is interested is unchanged. Graying is carried out on the image after semantic segmentation to obtain a surface gray level diagram of the aluminum material, and the schematic diagram is shown as follows. When the image was observed, two inconspicuous dirty points were found, which appeared as a highlight region. However, due to the interference of other abnormally high areas on the metal surface, defects cannot be accurately detected directly through threshold segmentation, and further discussion of how to denoise is needed, so that image enhancement is completed, and the defects are more obvious in the image. Thus, the process of collecting the surface image of the aluminum material and preprocessing is completed.
In the complex processing process flow of aluminum materials, the reasons for the dirty points are numerous, and the expression forms of the dirty points are different. For dirty points with larger contrast ratio with the metal surface in gray scale, the defect detection can be completed by threshold segmentation directly. But the dirty points on the surface caused by breakage have certain concealment and are complex in characteristics. Considering that the dirty point defect to be detected is not a completely highlight region, but a circle of low gray pixels surrounding the highlight region, an abnormal region is first determined by calculating a gray variance. In the abnormal region, combining the characteristics of gradient and gray level symbiosis to obtain the defect credibility of the abnormal region, and then denoising is selectively performed according to the credibility.
The edge of the abnormal region in the aluminum surface gray level diagram has obvious change in gradient, but because the to-be-detected dirty point defect image not only comprises an abnormal highlight region, but also has a circle of abnormal low pixels outside the highlight region. In order to obtain the whole region of the dirty point defect, the region grows on the suspected defect gray level map, and the inner edge and the outer edge are obtained through the mutation of gray level variance, so that the complete abnormal region is obtained.
Calculating the gradient of the target image by utilizing a Sobel operator, wherein the method comprises the following detailed steps of: firstly, inputting an aluminum profile surface gray level map, setting a convolution kernel size as 3*3, respectively convoluting with a corresponding weight matrix in the horizontal direction and the vertical direction to obtain a first derivative, and returning to two unidirectional gradient maps; and then, carrying out mixed weighting on the gradient graphs in the horizontal and vertical directions to obtain a gradient image. Dividing the gradient map by using an Ojin threshold method, and multiplying the division result by the original gray map as a mask to obtain a suspected defect gray map.
Setting sliding windows with preset sizes on the suspected defect gray level map, obtaining gray level maximum value points in each window with the preset sizes being 7*7, taking the gray level maximum value points as seed points for region growth, taking the seed points as the center, expanding the region,representing the number of expansions. In the expansion process, the gray variance of the corresponding region is calculated as follows:
logic: in the middle ofIs through->The number of pixels in the area is obtained after the sub-expansion, < >>Is the>Gray values of individual pixels. For the expanded region, the average gray level is calculated and then the variance of gray level is calculated>。/>The larger the fluctuation degree of the pixel gradation in the region is, the larger; and vice versa.
For a gray maximum point, a gray variance sequence of its expanded region can be obtained. Since the dirty points to be detected appear as: there is a circle of very low gray pixels surrounding the highlight region, with two edges, compared to the lower background pixels. Therefore, the mutation value is found in the gray sequence, the region edge corresponding to the first mutation is taken as the inner edge, and the region edge corresponding to the second mutation is taken as the outer edge. The outer edge is taken as the boundary of the abnormal region where dirty points may exist. Traversing all the maximum gray value points to obtain a plurality of abnormal areas in the suspected defect gray mapWherein->Is the total number of abnormal regions.
To this end, the process of determining an abnormal region where a defect may exist from the image gradient is completed.
Step S2, gradient abnormality of the abnormal region is obtained according to the average gradient of the edge of the abnormal region; setting a threshold value, obtaining gray level co-occurrence matrixes of each abnormal region in four directions, and obtaining point-to-gray level abnormality of the abnormal region based on point pairs with absolute values of gray level differences of point pairs in the gray level co-occurrence matrixes in each direction being larger than the threshold value; setting a threshold value, and obtaining a direction sequence based on the direction of a connecting line of pixel points in a point pair with an absolute value of a gray level difference of a point pair in a gray level co-occurrence matrix in each direction, wherein the connecting line is larger than the threshold value; the product of the number of elements in the direction sequence and the point-to-gray level abnormality is the gray level symbiotic abnormality of the abnormal region.
Since some noise regions may also have a feature of abrupt gray variance, the abnormal regions do not necessarily have defects, and thus it is also necessary to distinguish dirty points from noise by other properties. Considering that compared with a noise area affecting the judgment of the dirty point defect, the true dirty point defect area has pixels with extremely high gray values and pixels with extremely low gray values, has more complex textures, has a plurality of gray point pairs with larger differences, and has various corresponding gradient directions, and the defect credibility of each abnormal area is calculated through a gray co-occurrence matrix.
In acquiring abnormal regionIn this case, the gray scale of the true dirty point defect region is changed from inside to outside: from highlight to extreme dark to darker, the inner and outer edges are determined by abrupt changes in the gray variance sequence, but only qualitative discussion. Calculating the ratio of the average gradient of the pixels on the inner edge to the average gradient of the pixels on the outer edge to obtain +.>Gradient abnormality of the abnormality region->。/>The larger the gray scale change characteristic of the abnormal region is, the more similar to the dirty point region, and the more likely the abnormal region is the true dirty point region.
First, theAbnormal region->Setting the offset as 1, and calculating gray level co-occurrence matrixes in four directions of 0 degree, 45 degree, 90 degree and 135 degree. The elements in the matrix are gray point pairs +.>The number of co-occurrences in the specified direction and offset values.
Point-to-point pairsThe gray scale difference of (2) is->. Respectively constructing a gray difference histogram for point pairs in the gray co-occurrence matrix in four directions, wherein the horizontal axis is the gray difference of the point pairs in the gray co-occurrence matrix in the direction, and the vertical axis is the occurrence frequency of the difference. Threshold value of set point-to-gray difference +.>Will be different by more than->Is proposed +.>Is 100, calculates the abnormal region +.>Point-to-gray level anomalies->The formula is as follows:
logic in the abnormal regionIn (2) for gray level co-occurrence matrix in one direction, < ->Is the gray point pair differenceThe corresponding frequency of occurrence will be greater than +.>Is accumulated for the degree of difference. And then the accumulated values in the four directions are averaged to obtain the +.>Gray level abnormality of individual abnormality region +.>。/>The larger the difference in gray value of the gray point pair of the region is, the larger the difference in gray value is.
In a true dirty point defect region, the gray point pairs have large differences in values, and the directions of the large-gap point pairs are also more varied. Therefore, in the abnormal regionMid-statistical gray level difference->The connection line direction of the point pair of (2) to obtain a direction sequence, wherein the value range of the connection line direction of the point pair is [ -DEG C>. Examples: and counting all pixel point pairs with gray values of 10 and 120 in the abnormal area in the wiring direction, wherein the point pairs meeting the requirements obtained by the gray level co-occurrence matrix calculation in the 0-degree direction are (10, 120). Calculating the number of elements in the direction sequence +.>The larger the number of elements, the stronger the directional diversity. Dot-to-gray level abnormality obtained by combining>The gray level symbiotic abnormality of the abnormal region can be obtained from the two angles of the difference and the direction of the point pair +.>The calculation formula is as follows:
logic: a large number of dot pairs with large gray scale difference exist in the gray scale co-occurrence matrix of the dirty dot area, and the connecting line directions of the dot pairs with large difference are various. Abnormal gray levelAnd number of direction sequence elements->Multiplication to obtain the abnormality of gray level co-occurrence matrix from both value and direction>。/>The larger the gray point pair difference representing the region, the more likely it is that a truly dirty point region.
Step S3, obtaining defect credibility according to gradient abnormality and gray level symbiotic abnormality of the abnormal region; sequencing the defect credibility of all abnormal areas, and obtaining the lower quartile based on sequencing results; the defect area with the defect credibility greater than the lower quartile is a suspected dirty point defect area; filtering other areas except for the suspected dirty point defect area in the surface gray scale map to obtain a filtered image; and detecting the quality of the surface of the aluminum product according to the filtered image.
Complex gradient abnormalityAnd gray level symbiotic abnormality->Defect confidence level of obtaining abnormal region +.>The calculation formula is as follows:
logic: due to the dirty point defect regionHas abnormal expression in the gradient and gray level co-occurrence matrixThe larger the product is, the stronger the dirty point defect is represented by the abnormal region, and the more likely the abnormal region is a true dirty point defect region. Mapping the reciprocal of the product of the two to [0,1 ] by using an exponential function]Interval: the larger the product, the smaller the reciprocal, the mappedThe closer the value is to 1; the smaller the product, the larger the reciprocal>The closer to 0 the value of (c). The defect confidence level represents the probability that the abnormal region is a true dirty point defect, < >>The larger the abnormal region->The more likely a dirty point defect, the less likely a noisy region.
So far, discussing the gradient and gray level co-occurrence matrix characteristics of the abnormal region, respectively calculating the abnormality, and completing the process of comprehensively obtaining the credibility of the dirty points.
And denoising the image according to the credibility of the dirty points. The specific process comprises the following steps: respectively calculating the defect credibility of the abnormal region in the image to obtain a defect credibility sequence. Ordering the credibility of abnormal areas in the sequence, finding out the lower quartile, and reserving the abnormal areas with weights above the lower quartile, wherein the abnormal areas are very likely to be real dirty point defect areas. And denoising the rest area in the image based on a filter, and adopting average filtering. Noise which is easy to confuse dirty points is effectively eliminated, so that the edge of a non-defect abnormal area is blurred, and a filtered image is obtained.
So far, the process of denoising the image according to the reliability of the dirty points is completed.
The quantitative denoising is finished based on the credibility of the dirty point defects, and the image is enhanced essentially, so that the noise which can interfere with defect detection under the segmentation of a general threshold is removed, and the defects with high credibility are highlighted. And (3) re-carrying out threshold segmentation on the filtered image, so that the region where the dirty point defects are located can be more accurately positioned and marked, and two unobvious dirty point defects are detected. Although threshold segmentation can only obtain a highlight region in the dirty point defect in a damaged state, and cannot obtain a layer of low gray scale region on the outer side, the distinction from other abnormal highlight regions is completed by calculating the defect credibility, and the highlight region marked after segmentation is the region where the dirty point defect is located. Further, the position parameters of the region where the defect is located can be output, so that enterprises can know the surface state of the aluminum profile product.
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. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures 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.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (10)
1. The method for detecting the surface quality of the aluminum material is characterized by comprising the following steps:
collecting an image of the surface of the aluminum material and preprocessing to obtain a surface gray level image of the aluminum material; obtaining a suspected defect gray scale map in the surface gray scale map; obtaining at least two gray maximum points in the suspected defect gray map; performing region expansion at least twice by taking the gray maximum point as a center to obtain a region sequence corresponding to the gray maximum point; calculating the gray variance of each region of the region sequence corresponding to the gray maximum point to form a gray variance sequence; obtaining an abnormal region corresponding to a gray maximum point according to the edges of the regions corresponding to the first mutation point and the second mutation point in the gray variance sequence;
obtaining gradient abnormality of the abnormal region according to the average gradient of the edge of the abnormal region; setting a threshold value, obtaining gray level co-occurrence matrixes of each abnormal region in four directions, and obtaining point-to-gray level abnormality of the abnormal region based on point pairs with absolute values of gray level differences of point pairs in the gray level co-occurrence matrixes in each direction being larger than the threshold value; obtaining a direction sequence based on the direction of the connecting line of the pixel points in the point pair with the absolute value of the gray difference of the point pair larger than the threshold value in the gray level co-occurrence matrix in each direction; the product of the number of elements in the direction sequence and the point-to-gray level abnormality is the gray level symbiotic abnormality of the abnormal region;
obtaining defect credibility according to gradient abnormality and gray level symbiotic abnormality of an abnormal region; screening the abnormal region according to the defect reliability to obtain a suspected dirty point defect region; filtering other areas except for the suspected dirty point defect area in the surface gray scale map to obtain a filtered image; and detecting the quality of the surface of the aluminum product according to the filtered image.
2. The method for detecting the surface quality of the aluminum material according to claim 1, wherein the steps of collecting an image of the surface of the aluminum material and performing pretreatment to obtain a surface gray scale map of the aluminum material comprise: carrying out semantic segmentation on the image on the surface of the aluminum material, setting the pixel value of the background area to be 0, and keeping the pixel value of the aluminum material area unchanged to obtain the image after semantic segmentation; and carrying out graying treatment on the semantically segmented image to obtain a surface gray scale image.
3. The method for detecting the surface quality of aluminum materials according to claim 1, wherein the obtaining at least two gray maxima points in the gray map of suspected defects comprises:
setting a sliding window with a preset size, and when the sliding window slides on the suspected defect gray scale map, obtaining a pixel point with the maximum gray scale value in each sliding window as a gray scale maximum value point.
4. The method for detecting the surface quality of aluminum material according to claim 1, wherein the step of obtaining an abnormal region corresponding to a maximum gray value point from the edge of the region corresponding to the first and second abrupt points in the gray variance sequence comprises: the edge of the region in the region sequence corresponding to the first mutation point is the inner edge of the abnormal region, and the edge of the region in the region sequence corresponding to the second mutation point is the outer edge of the abnormal region.
5. The method for detecting the surface quality of an aluminum material according to claim 1, wherein the obtaining gradient abnormality of the abnormal region from the average gradient of the edge of the abnormal region comprises: the ratio of the average gradient value of the inner edge to the average gradient value of the outer edge of the abnormal region is the gradient abnormality of the abnormal region.
6. The method for detecting the surface quality of an aluminum material according to claim 1, wherein the point-to-gray scale abnormality is:
wherein,point-to-gray scale anomalies representing the s-th anomaly region; />Respectively representing gray level co-occurrence matrixes in four directions of an abnormal region; />Representing a threshold value; />A set of point pairs representing a gray level co-occurrence matrix of the s-th anomaly region; />An absolute value representing a difference in gray values of two pixel points in a pair of points; />The frequency of occurrence of the absolute value of the difference in gray value of two pixel points in a pair is represented.
7. The method for detecting the surface quality of the aluminum material according to claim 1, wherein the defect reliability degree is:
wherein,representing the defect credibility of the s-th abnormal region; />Gradient abnormality representing the s-th abnormality region; />Gray level co-occurrence abnormality representing the s-th abnormal region; />An exponential function based on a natural constant e is represented.
8. The method for detecting the surface quality of the aluminum material according to claim 1, wherein the detecting the surface quality of the aluminum material based on the filtered image comprises: and (3) re-carrying out threshold segmentation on the filtered image to obtain a highlight region in the dirty point defects in a damaged state, wherein the highlight region marked after segmentation is the region where the dirty point defects are located, and outputting the position parameters of the region where the defects are located.
9. The method for detecting the surface quality of the aluminum material according to claim 1, wherein the method for acquiring the suspected defect gray-scale map comprises the following steps:
obtaining a gradient image of the surface gray level map, and dividing the gradient image to obtain a mask image; and multiplying the mask image by the surface gray scale map to obtain a suspected defect gray scale map.
10. The method for detecting the surface quality of an aluminum material according to claim 1, wherein the method for obtaining the suspected dirty point defect region comprises:
sequencing the defect credibility of all abnormal areas, and obtaining the lower quartile based on sequencing results; the abnormal region with the defect reliability degree larger than the lower quartile is a suspected dirty point defect region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410121803.6A CN117649412B (en) | 2024-01-30 | 2024-01-30 | Aluminum material surface quality detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410121803.6A CN117649412B (en) | 2024-01-30 | 2024-01-30 | Aluminum material surface quality detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117649412A true CN117649412A (en) | 2024-03-05 |
| CN117649412B CN117649412B (en) | 2024-04-09 |
Family
ID=90045464
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410121803.6A Active CN117649412B (en) | 2024-01-30 | 2024-01-30 | Aluminum material surface quality detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117649412B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117951455A (en) * | 2024-03-22 | 2024-04-30 | 汶上义桥煤矿有限责任公司 | On-line monitoring method for operation faults of scraper conveyor |
| CN117953316A (en) * | 2024-03-27 | 2024-04-30 | 湖北楚天龙实业有限公司 | Image quality inspection method and system based on artificial intelligence |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3749726B1 (en) * | 2005-06-01 | 2006-03-01 | 株式会社ファースト | Low contrast defect inspection method under periodic noise, low contrast defect inspection method under repeated pattern |
| JP2011017609A (en) * | 2009-07-09 | 2011-01-27 | Panasonic Corp | Defect inspecting apparatus and defect inspecting method |
| WO2020248439A1 (en) * | 2019-06-11 | 2020-12-17 | 江苏农林职业技术学院 | Crown cap surface defect online inspection method employing image processing |
| CN112270658A (en) * | 2020-07-13 | 2021-01-26 | 安徽机电职业技术学院 | Elevator steel wire rope detection method based on machine vision |
| KR20220102506A (en) * | 2021-01-13 | 2022-07-20 | 한국전력공사 | Apparatus and method for detecting of faults in surfaces of metal tube |
| 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 |
| CN114972326A (en) * | 2022-07-20 | 2022-08-30 | 南通鼎彩新材料科技有限公司 | Defective product identification method for heat-shrinkable tube expanding process |
| CN115115642A (en) * | 2022-08-30 | 2022-09-27 | 启东万惠机械制造有限公司 | Strip steel scab defect detection method based on image processing |
| CN115272316A (en) * | 2022-09-27 | 2022-11-01 | 山东华太新能源电池有限公司 | Intelligent detection method for welding quality of battery cover based on computer vision |
| CN115311304A (en) * | 2022-10-12 | 2022-11-08 | 江苏明锋食品有限公司 | Iron plate corrosion defect detection method |
| CN115345885A (en) * | 2022-10-19 | 2022-11-15 | 南通鹏宝运动用品有限公司 | Method for detecting appearance quality of metal fitness equipment |
| CN115661136A (en) * | 2022-12-12 | 2023-01-31 | 深圳宝铭微电子有限公司 | Semiconductor defect detection method for silicon carbide material |
| WO2023077404A1 (en) * | 2021-11-05 | 2023-05-11 | 宁德时代新能源科技股份有限公司 | Defect detection method, apparatus and system |
| WO2023134792A2 (en) * | 2022-12-15 | 2023-07-20 | 苏州迈创信息技术有限公司 | Led lamp wick defect detection method |
| CN116862908A (en) * | 2023-09-01 | 2023-10-10 | 菏泽学院 | Metal gear meshing defect detection method based on image processing |
-
2024
- 2024-01-30 CN CN202410121803.6A patent/CN117649412B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3749726B1 (en) * | 2005-06-01 | 2006-03-01 | 株式会社ファースト | Low contrast defect inspection method under periodic noise, low contrast defect inspection method under repeated pattern |
| JP2011017609A (en) * | 2009-07-09 | 2011-01-27 | Panasonic Corp | Defect inspecting apparatus and defect inspecting method |
| WO2020248439A1 (en) * | 2019-06-11 | 2020-12-17 | 江苏农林职业技术学院 | Crown cap surface defect online inspection method employing image processing |
| CN112270658A (en) * | 2020-07-13 | 2021-01-26 | 安徽机电职业技术学院 | Elevator steel wire rope detection method based on machine vision |
| KR20220102506A (en) * | 2021-01-13 | 2022-07-20 | 한국전력공사 | Apparatus and method for detecting of faults in surfaces of metal tube |
| WO2023077404A1 (en) * | 2021-11-05 | 2023-05-11 | 宁德时代新能源科技股份有限公司 | Defect detection method, apparatus and system |
| 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 |
| CN114972326A (en) * | 2022-07-20 | 2022-08-30 | 南通鼎彩新材料科技有限公司 | Defective product identification method for heat-shrinkable tube expanding process |
| CN115115642A (en) * | 2022-08-30 | 2022-09-27 | 启东万惠机械制造有限公司 | Strip steel scab defect detection method based on image processing |
| CN115272316A (en) * | 2022-09-27 | 2022-11-01 | 山东华太新能源电池有限公司 | Intelligent detection method for welding quality of battery cover based on computer vision |
| CN115311304A (en) * | 2022-10-12 | 2022-11-08 | 江苏明锋食品有限公司 | Iron plate corrosion defect detection method |
| CN115345885A (en) * | 2022-10-19 | 2022-11-15 | 南通鹏宝运动用品有限公司 | Method for detecting appearance quality of metal fitness equipment |
| CN115661136A (en) * | 2022-12-12 | 2023-01-31 | 深圳宝铭微电子有限公司 | Semiconductor defect detection method for silicon carbide material |
| WO2023134792A2 (en) * | 2022-12-15 | 2023-07-20 | 苏州迈创信息技术有限公司 | Led lamp wick defect detection method |
| CN116862908A (en) * | 2023-09-01 | 2023-10-10 | 菏泽学院 | Metal gear meshing defect detection method based on image processing |
Non-Patent Citations (8)
| Title |
|---|
| 俞翔;胡志强;: "基于局部纹理特征的钢卷侧面缺陷检测方法研究", 计算机工程与设计, no. 24, 28 December 2009 (2009-12-28) * |
| 吴彬彬;汤勃;孔建益;王兴东;: "MAS小波的钢板表面缺陷边缘检测的研究", 机械设计与制造, no. 05, 8 May 2015 (2015-05-08) * |
| 文生平;李超贤;: "基于Gaussian-yolov3的铝型材表面缺陷检测", 计算机测量与控制, no. 09, 25 September 2020 (2020-09-25) * |
| 甘胜丰;雷维新;邓芳;袁荣奇;: "钢材表面缺陷图像感兴趣区域提取方法", 机械设计与制造, no. 01, 8 January 2017 (2017-01-08) * |
| 白雪冰;王科俊;邹丽晖;: "基于二维阈值向量的木材表面缺陷分割方法", 东北林业大学学报, no. 09, 25 September 2008 (2008-09-25) * |
| 邢茹;李晓欣;赵建军;: "基于激光声磁技术的钢轨踏面缺陷检测", 激光杂志, no. 09, 25 September 2018 (2018-09-25) * |
| 陈后金;许文达;郝晓莉;: "基于灰度-梯度共生矩阵的钢轨表面缺陷检测方法", 北京交通大学学报, no. 02, 15 April 2015 (2015-04-15) * |
| 黄战华;刘正;朱猛;蔡怀宇;张尹馨;: "基于统计特征的轮胎纹理缺陷在线检测", 光学技术, no. 01, 20 January 2009 (2009-01-20) * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117951455A (en) * | 2024-03-22 | 2024-04-30 | 汶上义桥煤矿有限责任公司 | On-line monitoring method for operation faults of scraper conveyor |
| CN117951455B (en) * | 2024-03-22 | 2024-06-07 | 汶上义桥煤矿有限责任公司 | On-line monitoring method for operation faults of scraper conveyor |
| CN117953316A (en) * | 2024-03-27 | 2024-04-30 | 湖北楚天龙实业有限公司 | Image quality inspection method and system based on artificial intelligence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117649412B (en) | 2024-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111325721A (en) | Gas leakage detection method and system based on infrared thermal imaging | |
| CN115829883B (en) | Surface image denoising method for special-shaped metal structural member | |
| CN117649412B (en) | Aluminum material surface quality detection method | |
| CN113505865B (en) | Sheet surface defect image recognition processing method based on convolutional neural network | |
| CN114638827B (en) | Visual detection method and device for impurities of lubricating oil machinery | |
| CN112184693B (en) | Intelligent detection method for welding line defects of ray industrial negative film | |
| CN117351019B (en) | Welding defect detection method | |
| CN115063409B (en) | Method and system for detecting surface material of mechanical cutter | |
| CN109239073B (en) | Surface defect detection method for automobile body | |
| CN116758083A (en) | Quick detection method for metal wash basin defects based on computer vision | |
| CN116091504B (en) | Connecting pipe connector quality detection method based on image processing | |
| CN113362306A (en) | Packaged chip defect detection method based on deep learning | |
| CN111667470B (en) | Industrial pipeline flaw detection inner wall detection method based on digital image | |
| CN110648330B (en) | Defect detection method for camera glass | |
| CN115841434A (en) | Infrared image enhancement method for gas concentration analysis | |
| CN117437219B (en) | Magnetic glue inductance irregular appearance defect detection method based on computer vision | |
| CN116152242B (en) | Visual detection system of natural leather defect for basketball | |
| CN107341793A (en) | A kind of target surface image processing method and device | |
| CN116309416A (en) | Visual inspection method for surface of part | |
| CN117689662B (en) | Visual detection method and system for welding quality of heat exchanger tube head | |
| CN115330770B (en) | Cloth area type defect identification method | |
| CN111192261A (en) | Method for identifying lens defect types | |
| CN114486916A (en) | Mobile phone glass cover plate defect detection method based on machine vision | |
| CN114998186A (en) | Image processing-based method and system for detecting surface scab defect of copper starting sheet | |
| CN110390649B (en) | Method for reducing noise of oil and gas pipeline weld image |
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 |