CN119067923B - Intelligent defect detection method for milk outer package - Google Patents
Intelligent defect detection method for milk outer package Download PDFInfo
- Publication number
- CN119067923B CN119067923B CN202411083204.6A CN202411083204A CN119067923B CN 119067923 B CN119067923 B CN 119067923B CN 202411083204 A CN202411083204 A CN 202411083204A CN 119067923 B CN119067923 B CN 119067923B
- Authority
- CN
- China
- Prior art keywords
- pixels
- bottle bottom
- area
- pixel
- grayscale
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 108
- 239000008267 milk Substances 0.000 title claims abstract description 27
- 210000004080 milk Anatomy 0.000 title claims abstract description 27
- 235000013336 milk Nutrition 0.000 title claims abstract description 27
- 238000001514 detection method Methods 0.000 title claims abstract description 18
- 230000002950 deficient Effects 0.000 claims abstract description 20
- 238000012216 screening Methods 0.000 claims abstract description 18
- 238000004806 packaging method and process Methods 0.000 claims abstract description 14
- 238000009826 distribution Methods 0.000 claims abstract description 11
- 239000011521 glass Substances 0.000 claims description 22
- 238000000034 method Methods 0.000 claims description 16
- 238000010606 normalization Methods 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 13
- 230000011218 segmentation Effects 0.000 claims description 11
- 230000000877 morphologic effect Effects 0.000 claims description 8
- 230000005484 gravity Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 6
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 5
- 238000003708 edge detection Methods 0.000 description 5
- 238000005286 illumination Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000003709 image segmentation Methods 0.000 description 3
- 230000001788 irregular Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 239000000796 flavoring agent Substances 0.000 description 1
- 235000019634 flavors Nutrition 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000000813 microbial effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/04—Context-preserving transformations, e.g. by using an importance map
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20036—Morphological image processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
本发明涉及图像处理技术领域，具体涉及一种用于牛奶外包装的缺陷智能化检测方法，包括：获取瓶底灰度图像并得到疑似区域，计算瓶底灰度图像的每个像素点的8邻域灰度均匀度，进而筛选出目标区域，在每个目标区域内筛选出若干个高亮点，根据高亮点及其周围像素点的8邻域灰度均匀度和梯度，在所有目标区域中筛选出若干个缺陷区域，根据缺陷区域特征寻找漏选的缺陷区域，根据所有缺陷区域在瓶底灰度图像中的位置分布，判断瓶底的外包装是否为合格外包装。本发明根据区域间的灰度差异及边缘特征进行筛选干扰信息，最终确定缺陷区域，有效提高了现有技术识别瓶装牛奶外包装缺陷的准确性和效率，减少了细节信息的缺失。
The present invention relates to the field of image processing technology, and in particular to an intelligent defect detection method for milk outer packaging, comprising: obtaining a grayscale image of a bottle bottom and obtaining a suspected area, calculating the 8-neighborhood grayscale uniformity of each pixel point of the bottle bottom grayscale image, and then screening out a target area, screening out a number of highlight points in each target area, screening out a number of defective areas in all target areas according to the 8-neighborhood grayscale uniformity and gradient of the highlight points and the surrounding pixels, finding missed defective areas according to defective area characteristics, and judging whether the outer packaging of the bottle bottom is a qualified outer packaging according to the position distribution of all defective areas in the bottle bottom grayscale image. The present invention screens interference information according to the grayscale difference and edge characteristics between regions, and finally determines the defective area, effectively improving the accuracy and efficiency of identifying defects in the outer packaging of bottled milk in the prior art, and reducing the loss of detailed information.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an intelligent defect detection method for milk outer package.
Background
The glass bottle has good chemical stability and good barrier effect, can not react with milk, is favorable for avoiding pollution of air to the milk, prevents diffusion of flavor substances, and effectively maintains purity and nutrition components of the milk, but is fragile, notches are easy to appear, and if the milk is packaged by the glass bottle with the notches, potential safety hazards are easy to appear, so that the milk is directly exposed to the air, and further microbial pollution occurs.
The glass material has poor adaptability to temperature, is fragile and has poor impact resistance, so the glass bottle has the most common defects of bottom abrasion or notch caused by unsuitable temperature in the process of gob production or collision in the process of transportation and storage, and the defect detection of the milk outer package is mostly carried out by manual detection or image threshold segmentation and edge detection at present, but because a large amount of textures exist in the milk glass bottle, the manual detection and image segmentation are interfered, and the accuracy of the defect detection of the milk outer package is possibly reduced.
Disclosure of Invention
The invention provides an intelligent defect detection method for milk outer package, which aims to solve the existing problems.
The intelligent defect detection method for milk outer package adopts the following technical scheme:
An embodiment of the invention provides an intelligent defect detection method for milk outer package, which comprises the following steps:
The method comprises the steps of acquiring a bottle bottom gray level image, carrying out threshold segmentation on the bottle bottom gray level image to obtain suspected areas, and acquiring an edge chain code sequence of each suspected area;
According to the 8 neighborhood gray level uniformity of all the pixel points in the bottle bottom gray level image and the edge chain code sequence of each suspected region, screening out a target region from all the suspected regions of the bottle bottom gray level image;
screening a plurality of high-brightness points on all edge lines in each target area according to the gray values of all pixel points in each target area, and screening a plurality of defect areas in all target areas according to the 8 neighborhood gray uniformity and gradient of the high-brightness points and the surrounding pixel points;
carrying out morphological operation on all the defect areas to obtain a plurality of possible connected domains;
and judging whether the outer package of the bottle bottom is a qualified outer package according to the position distribution of all the defect areas in the gray level image of the bottle bottom.
Further, the specific calculation formula for obtaining the 8 neighborhood gray uniformity of each pixel point of the bottle bottom gray image according to the gradient amplitude values and gray values of all the pixel points in the bottle bottom gray image is as follows:
Wherein q i represents the 8 neighborhood gray scale uniformity of the ith pixel point of the bottle bottom gray scale image, t i represents the sum of the gradient magnitudes of all pixel points in the 8 neighborhood of the ith pixel point of the bottle bottom gray scale image, and h i represents the average value of the gray scale values of all pixel points in the 8 neighborhood of the ith pixel point of the bottle bottom gray scale image; and s i represents the variance of the gray values of all pixels in the 8 neighborhood of the ith pixel of the bottle bottom gray image.
Further, the step of screening the target area from all the suspected areas of the bottle bottom gray level image according to the 8 neighborhood gray level uniformity of all the pixel points in the bottle bottom gray level image and the edge chain code sequence of each suspected area comprises the following specific steps:
obtaining the degree of conforming to the defect characteristics of each suspected region according to the 8 neighborhood gray level uniformity of all pixel points in the bottle bottom gray level image;
And (3) in all suspected areas of the bottle bottom gray level image, marking the suspected areas which are in accordance with the defect characteristics and have the degree larger than the preset defect characteristic degree threshold as target areas.
Further, according to the 8 neighborhood gray scale uniformity of all pixel points in the bottle bottom gray scale image, a specific calculation formula for obtaining the degree that each suspected region accords with the defect feature is:
Wherein p j represents the degree that the jth suspected region accords with the defect characteristic, v j represents the number of pixel points in the jth suspected region, and q j,a represents the 8-neighborhood gray scale uniformity of the a-th pixel point in the jth suspected region; The average value of 8 neighborhood gray level uniformity of all pixel points in the bottle bottom gray level image is represented, l j represents the number of the maximum continuous repeated chain codes in the edge chain code sequence of the jth suspected region, and norm () is a linear normalization function.
Further, the step of screening a plurality of highlight points on all edge lines in each target area according to the gray values of all pixel points in each target area includes the following specific steps:
And in the 8 adjacent areas of the y-th edge pixel points on all edge lines of the x-th target area, when the gray level value of the y-th edge pixel point is larger than the gray level value of all pixel points in the 8 adjacent areas of the y-th edge pixel point and the number of the pixel points with the gray level value larger than the average gray level value in the 8 adjacent areas of the y-th edge pixel point is larger than or equal to the preset number of the pixel points, marking the y-th edge pixel point as a highlight point.
Further, the step of screening a plurality of defect areas from all the target areas according to the 8 neighborhood gray scale uniformity and gradient of the highlight points and surrounding pixel points comprises the following specific steps:
Acquiring normal lines of each highlight point on all edge lines in an x-th target area, and marking the front b adjacent pixel points on two sides of each highlight point as pixel points to be detected of the corresponding highlight point of the x-th target area on the normal line of each highlight point, wherein b is the preset number of pixel points;
Obtaining the degree that each pixel point to be detected of each highlight point of each target area accords with the pixel point of the text area according to the 8 neighborhood gray level uniformity and the gradient direction vector of all the pixel points to be detected of each target area;
among all the pixel points to be detected in each target area, the pixel points to be detected, which accord with the pixel points of the text area and have the degree of being greater than the preset effective pixel point threshold value, are marked as effective pixel points;
obtaining the degree of conforming to the text region of each target region according to the degree of conforming to the text region pixel points of all the pixel points to be detected in each target region and the number of effective pixel points;
and (3) in all target areas, marking the target areas which accord with the text areas and have the degree smaller than the preset defect area threshold value as defect areas.
Further, according to the 8-neighborhood gray uniformity and the gradient direction vector of all the pixels to be detected in each target area, a specific calculation formula for obtaining the degree to which each pixel to be detected of each highlight point in each target area accords with the pixels of the text area is:
In the formula, Representing the degree to which the mth pixel point to be detected of the kth highlight point of the a-th target area accords with the pixel point of the text area; Representing 8 neighborhood gray scale uniformity of an mth pixel to be detected of a kth highlight point of an a-th target area; The m-th pixel point to be detected of the kth highlight point of the a-th target area is represented, and the 8-neighborhood gray scale uniformity of the pixel point to be detected which is symmetrical about the highlight point on the corresponding normal line is represented; a gradient of an mth pixel to be detected representing a kth highlight of the a-th target area; A gradient of an mth pixel to be detected representing a (k+1) th highlight of the a-th target region; Representation of AndThe degrees of the included angle are formed, the absolute value function is the I, and the norm () is the linear normalization function.
Further, according to the degree that all the pixels to be detected in each target area meet the pixels of the text area and the number of the effective pixels, a specific calculation formula for obtaining the degree that each target area meets the text area is:
in the formula, w n represents the degree that the nth target area accords with the text area, c n represents the number of all pixel points to be detected in the nth target area, u n,d represents the degree that the d pixel point to be detected in the nth target area accords with the pixel points of the text area, c ′ n represents the number of all effective pixel points in the nth target area, and norm () is a linear normalization function.
Further, the screening of the plurality of defect areas from the plurality of possible connected areas includes the following steps:
Performing circle fitting on the gray level image of the bottle bottom to obtain a circle with the largest radius in the fitting result, and marking the circle as the outermost circle curve of the bottle bottom;
marking all 8 neighborhood pixel points of all pixel points of the bottle bottom outermost circular curve and pixel points in the intersection of all pixel points of each possible connected domain as pixel points of each possible connected domain on the bottle bottom outermost circular curve;
Marking the linear normalization value of the quotient of the number of the pixel points of each possible connected domain on the outermost circular curve of the bottle bottom and the average value of the shortest distance from all edge pixel points in each possible connected domain to the outermost circular curve of the bottle bottom as the degree that each possible connected domain accords with the self texture of the glass bottle;
and (3) marking the possible connected domains which accord with the texture of the glass bottle and are larger than the preset defect area threshold value as defect areas in all the possible connected domains.
Further, according to the position distribution of all the defect areas in the grayscale image of the bottle bottom, judging whether the outer package of the bottle bottom is a qualified outer package or not, comprising the following specific steps:
In the gray level image of the bottle bottom, the product of the shortest distance from the center of gravity of each defective area to the center of the circle of the outermost circular curve of the bottle bottom and the number of all pixel points in the defective area is recorded as the severity of each defective area;
Carrying out linear normalization processing on the severity of each defect area to obtain the normalized severity of each defect area;
Marking the average value of the normalized severity of all the defect areas as the defect degree of the detected bottle bottom;
and marking the bottom of the outer packaging bottle with the defect degree smaller than the preset defect area threshold as a qualified outer packaging.
The technical scheme of the invention has the beneficial effects that:
The embodiment of the invention comprises the steps of acquiring a bottle bottom gray level image, carrying out threshold segmentation on the bottle bottom gray level image to obtain suspected areas, acquiring an edge chain code sequence of each suspected area to provide a data basis for subsequent analysis and processing, obtaining 8 neighborhood gray level uniformity of each pixel point of the bottle bottom gray level image according to gradient amplitude values and gray level values of all pixel points in the bottle bottom gray level image, screening out target areas in all suspected areas of the bottle bottom gray level image according to the 8 neighborhood gray level uniformity of all pixel points and the edge chain code sequence of each suspected area, combining the 8 neighborhood gray level uniformity and edge information, accurately positioning the areas possibly with defects, further reducing the required range, screening out a plurality of high bright points on all edge lines in each target area according to gray level values of all pixel points in each target area, further screening out a plurality of defect areas in all target areas according to the 8 neighborhood gray level uniformity and gradient of the high bright points and surrounding pixel points, ensuring that the real defect areas are found out, carrying out morphological operation on all the defect areas, obtaining a plurality of possible connected areas, and finding out a plurality of possible connected defect areas in an outer package according to the position of the bottle bottom defect, and judging whether the obtained is connected defect areas are connected in an outer package. The method and the device screen interference information according to the gray level difference and the edge characteristics among the areas, and finally determine the defect area, thereby effectively improving the accuracy and the efficiency of identifying the defects of the bottled milk outer package in the prior art and reducing the loss of detail information.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of steps of an intelligent defect detection method for milk overwrapping according to the present invention;
FIG. 2 is an exemplary gray scale image of a bottle bottom provided in accordance with an embodiment of the present invention;
FIG. 3 is a graph showing the segmentation of an exemplary bottom gray image according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a text on the bottom of a bottle according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a threshold segmentation result of a defective area according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a defect area edge detection result according to an embodiment of the present invention;
FIG. 7 is a schematic diagram of a normal region sharing a boundary with a defective region 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 the intelligent detection method for the defects of the milk outer package according to the invention, which is provided by 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 invention provides a specific scheme of an intelligent defect detection method for milk outer package, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of steps of a method for intelligent detection of defects in milk overwrap according to one embodiment of the present invention is shown, the method comprising the steps of:
Step S001, acquiring bottle bottom gray level images, performing threshold segmentation on the bottle bottom gray level images to obtain suspected areas, and acquiring edge chain code sequences of each suspected area.
Acquiring bottle bottom images by using cameras arranged on a production line in advance;
and (5) graying the bottle bottom image to obtain a bottle bottom gray image.
It should be noted that, the glass bottle to be detected is transported by a conveyor belt, a light source and a camera are arranged at a fixed position, a bottle bottom image is obtained and gray-scale processing is performed, and an example bottle bottom gray-scale image is shown in fig. 2.
Calculating the bottle bottom gray level image by using an Ojin algorithm to obtain an optimal segmentation threshold value and a bottle bottom gray level image segmentation result, wherein FIG. 3 shows an example bottle bottom gray level image segmentation result, and the Ojin algorithm is a known technology;
and marking a communication domain formed by continuous adjacent pixel points as a suspected region in all pixel points with gray values larger than an optimal segmentation threshold value in the bottle bottom gray image, thereby obtaining a plurality of suspected regions.
The edge-chain code sequence of each suspected region is obtained using the Freeman chain code algorithm, which is well known.
Step S002, obtaining 8 neighborhood gray scale uniformity of each pixel point of the bottle bottom gray scale image according to the gradient amplitude values and gray scale values of all pixel points in the bottle bottom gray scale image, and screening out a target area from all suspected areas of the bottle bottom gray scale image according to the 8 neighborhood gray scale uniformity of all pixel points in the bottle bottom gray scale image and the edge chain code sequence of each suspected area.
Because the notch inside of the glass bottle is uneven, a plurality of tiny cut surfaces exist on the surface of the notch, the capability of reflecting light by the cut surfaces is strong, but the distribution of the cut surfaces inside the notch is irregular, so that the refraction angles of light are changeable and irregular, the defect area is brighter than the imaging of the normal area and the internal textures are more under the influence of illumination, and the notch surface of the glass material is often uneven (the notch is often formed by damaging the bottle body due to external force, a plurality of tiny cracks are formed along the notch direction in the notch forming process due to the specificity of the glass material), the notch is uneven in surface, the edge is rough, the notch is irregular in shape, and the normal area caused by light reflection is often smooth in edge and regular in shape.
According to the gradient amplitude values and the gray values of all the pixel points in the bottle bottom gray image, the corresponding calculation formula for obtaining the 8 neighborhood gray uniformity of each pixel point of the bottle bottom gray image is as follows:
Wherein q i represents the 8 neighborhood gray scale uniformity of the ith pixel point of the bottle bottom gray scale image, t i represents the sum of the gradient magnitudes of all pixel points in the 8 neighborhood of the ith pixel point of the bottle bottom gray scale image, and h i represents the average value of the gray scale values of all pixel points in the 8 neighborhood of the ith pixel point of the bottle bottom gray scale image; and s i represents the variance of the gray values of all pixels in the 8 neighborhood of the ith pixel of the bottle bottom gray image.
It should be noted that, in this embodiment, the gradient amplitude of the pixel point is obtained by using a sobel operator, which is a known technique, wherein the larger the sum t i of the gradient amplitudes of all pixel points in the 8 neighborhood of the ith pixel point is, the more disordered the texture distribution is; The larger the variance s i of the gray values of all the pixels in the 8 neighborhood of the ith pixel point is, the larger the gray value of the 8 neighborhood of the ith pixel point relative to the whole image is, namely the higher the brightness is, and the larger the gray change degree of all the pixels in the 8 neighborhood of the ith pixel point is.
According to the 8 neighborhood gray level uniformity of all pixel points in the bottle bottom gray level image, the corresponding calculation formula for obtaining the degree that each suspected region accords with the defect characteristics is:
Wherein p j represents the degree that the jth suspected region accords with the defect characteristic, v j represents the number of pixel points in the jth suspected region, and q j,a represents the 8-neighborhood gray scale uniformity of the a-th pixel point in the jth suspected region; The average value of 8 neighborhood gray level uniformity of all pixel points in the bottle bottom gray level image is represented, l j represents the number of the maximum continuous repeated chain codes in the edge chain code sequence of the jth suspected region, norm () is a linear normalization function, and the data value is normalized to be within the [0,1] interval.
The threshold value of the degree of the defect feature preset in this embodiment is 0.5, which is described as an example, and other values may be set in other embodiments, which is not limited in this embodiment.
And (3) in all suspected areas of the bottle bottom gray level image, marking the suspected areas which are in accordance with the defect characteristics and have the degree larger than the preset defect characteristic degree threshold as target areas.
Step S003, screening out a plurality of high-brightness points on all edge lines in each target area according to gray values of all pixel points in each target area, and screening out a plurality of defect areas in all target areas according to 8-neighborhood gray uniformity and gradient of the high-brightness points and surrounding pixel points.
The background of the glass bottle possibly has a trademark, characters and other areas belonging to the texture of the bottle body, the figure 4 shows a character schematic diagram of the bottle bottom, the form presented in the photo possibly interferes with the identification of a defect area due to the influence of shooting angle and illumination, the defect area is identified by morphological characteristics of the two areas, the raised glass bottle character area is taken as an example, the highest pixel point of the raised part has the highest brightness relative to other pixel points in the area under the influence of illumination due to the raised character area and regular shape, the connecting highlight points form a highlight line, the gray values of the pixel points on two sides of the highlight line are gradually reduced along the direction perpendicular to the highlight line with the highlight line as a starting point, the more uneven gray value distribution is, namely the 8 neighborhood gray uniformity of the pixel point is larger, and the 8 neighborhood gray uniformity difference of the pixel point on two sides of the highlight line is smaller due to the relatively regular shape of the character raised, the smaller pixel point is the 8 neighborhood uniformity of the pixel point is the smaller, and the pixel point in the gradient direction of the two adjacent pixel points on the same side along the extending direction of the highlight line is smaller, and the pixel point in the gradient direction is more than the regular pixel point is the more regular.
The number of preset pixels in this embodiment is b=5, which is described as an example, and other values may be set in other embodiments, which is not limited in this embodiment;
Performing edge detection on each target area by using a Canny edge detection algorithm to obtain a plurality of edge lines, wherein the Canny edge detection algorithm is a public-work technology;
In the 8 adjacent areas of the y-th edge pixel point on all edge lines of the x-th target area, when the gray value of the y-th edge pixel point is larger than the gray value of all pixel points in the 8 adjacent areas of the y-th edge pixel point and the number of the pixel points with the gray value larger than the average gray value in the 8 adjacent areas of the y-th edge pixel point is larger than or equal to the preset number of the pixel points, marking the y-th edge pixel point as a highlight point;
It should be noted that, the highlight line formed by the highlight points may be a plurality of discontinuous curve segments, and the analysis shows that, with the pixel point on the curve segment as a starting point, the gray level change of the pixel point gradually decreases to two sides along the direction perpendicular to the tangent line of the point, and the closer the distance from the pixel point on two sides to the highlight line is, the closer the 8 neighborhood gray level uniformity of the pixel point is.
Acquiring normal lines of each highlight point on all edge lines in an x-th target area, and marking the front b adjacent pixel points on two sides of each highlight point as pixel points to be detected of the corresponding highlight point of the x-th target area on the normal line of each highlight point, wherein b is the preset number of pixel points;
it should be noted that, when the number of pixels on any side of the highlight point is less than b on the normal line of any highlight point in the x-th target area, the number of pixels with the smallest number in two sides is taken as b, and the pixel to be detected is obtained.
According to the 8-neighborhood gray level uniformity and the gradient direction vector of all the pixel points to be detected of each target area, the corresponding calculation formula for obtaining the degree that each pixel point to be detected of each highlight point of each target area accords with the pixel point of the text area is:
In the formula, Representing the degree to which the mth pixel point to be detected of the kth highlight point of the a-th target area accords with the pixel point of the text area; Representing 8 neighborhood gray scale uniformity of an mth pixel to be detected of a kth highlight point of an a-th target area; The m-th pixel point to be detected of the kth highlight point of the a-th target area is represented, and the 8-neighborhood gray scale uniformity of the pixel point to be detected which is symmetrical about the highlight point on the corresponding normal line is represented; a gradient of an mth pixel to be detected representing a kth highlight of the a-th target area; A gradient of an mth pixel to be detected representing a (k+1) th highlight of the a-th target region; Representation of AndThe degrees of the included angle are formed, the absolute value function is the absolute value, the norm () is the linear normalization function, and the data value is normalized to be within the interval of 0, 1.
It is to be noted that,The gray level similarity degree of the mth pixel point to be detected of the kth highlight point of the a-th target area and the pixel point to be detected symmetrical on the corresponding normal line with respect to the highlight point is shown, the larger the value is, the larger the similarity degree is, and the more the pixel point to be detected accords with the characteristics of the pixel points of the character area; Representation of AndThe smaller the included angle is, the more consistent the gray level change is, the more consistent the distribution rule of the pixels in the text area is, whenIn the absence, will thetaAnd setting the formula to be 1, and ensuring that the formula is established.
The preset effective pixel threshold value of the embodiment is 0.7, which is described as an example, and other values may be set in other embodiments, which is not limited in the embodiment;
among all the pixel points to be detected in each target area, the pixel points to be detected, which accord with the pixel points of the text area and have the degree of being greater than the preset effective pixel point threshold value, are marked as effective pixel points;
According to the degree that all the pixel points to be detected in each target area accord with the pixel points of the text area and the number of the effective pixel points, the corresponding calculation formula of the degree that each target area accords with the text area is obtained as follows:
Wherein w n represents the degree that the nth target area accords with the text area, c n represents the number of all pixel points to be detected in the nth target area, u n,d represents the degree that the d pixel point to be detected in the nth target area accords with the pixel points of the text area, c ′ n represents the number of all effective pixel points in the nth target area, norm () is a linear normalization function, and the data value is normalized to be within the [0,1] interval.
It is to be noted that,The average value of the degree that all the pixel points to be detected in the nth target area accord with the pixel points of the text area is represented; the ratio of the number of all effective pixel points to the number of pixel points to be detected in the nth target area is shown, and the greater the ratio is, the greater the degree that the nth target area accords with the text area is.
The preset defect area threshold value in this embodiment is 0.5, which is described as an example, and other values may be set in other embodiments, which is not limited in this embodiment;
and (3) in all target areas, marking the target areas which accord with the text areas and have the degree smaller than the preset defect area threshold value as defect areas.
Step S004, morphological operation is carried out on all the defect areas to obtain a plurality of possible connected domains, and a plurality of defect areas are screened out from the plurality of possible connected domains.
In the process of collecting images, glass can generate refraction under a light source, so that the situation that an actual defect part is missed in an area obtained by threshold segmentation in the current scene can occur, and due to the influence of illumination, a glass notch can be high in edge brightness, and the situation that a concave area in the notch is low in brightness, so that a suspected area obtained by threshold segmentation is incomplete as shown in fig. 5.
Performing morphological expansion operation on all the defect areas to obtain expansion areas, wherein the morphological expansion operation is a known technology;
In the bottle bottom gray level image, the gray level value of the pixel point in the defect area is set to be 1, the gray level value of the pixel point in the non-defect area is set to be 0, a binary image is obtained, the binary image is subjected to morphological expansion operation, the expanded binary image is obtained, and the area with the gray level value of the pixel point of 1 in the expanded binary image corresponds to the area in the bottle bottom gray level image and is the expansion area.
The areas except all the defect areas in the expansion area are marked as possible areas;
In the possible region, a connected region formed by consecutive adjacent pixel points is referred to as a possible connected region.
It should be noted that, the possible connected domain obtained by the subtraction may be the defect region shown in fig. 6, or may be a region sharing a boundary with the defect region shown in fig. 7, and in order to determine the defect region, the edge features of the possible connected domain are analyzed to determine the possibility that the possible connected domain belongs to the defect region.
Because the area sharing the boundary with the defect area is a normal area, a plurality of pixel points conforming to the self texture characteristics of the glass bottle are necessarily arranged on the edge of the area, and the degree that the edge of the possibly connected area conforms to the self texture of the glass bottle is judged by curve fitting.
Performing circle fitting on the bottle bottom gray level image by using a least square method to obtain a circle with the largest radius in a fitting result, and marking the circle as the outermost circle curve of the bottle bottom;
marking all 8 neighborhood pixel points of all pixel points of the bottle bottom outermost circular curve and pixel points in the intersection of all pixel points of each possible connected domain as pixel points of each possible connected domain on the bottle bottom outermost circular curve;
According to the pixel points of each possible communicating domain on the outermost circular curve of the bottle bottom and the distances from all the edge pixel points to the outermost circular curve of the bottle bottom, the corresponding calculation formula of the degree that each possible communicating domain accords with the self texture of the glass bottle is obtained as follows:
Wherein r e represents the degree to which the e-th possible connected domain conforms to the texture of the glass bottle itself; Representing the average value of the shortest distance from all edge pixel points in the e possible connected domain to the outermost circular curve of the bottle bottom, f e representing the number of pixel points of the e possible connected domain on the outermost circular curve of the bottle bottom, and norm () being a linear normalization function for normalizing the data value to be within the [0,1] interval.
And (3) marking the possible connected domains which accord with the texture of the glass bottle and are larger than the preset defect area threshold value as defect areas in all the possible connected domains.
And S005, judging whether the outer package of the bottle bottom is a qualified outer package according to the position distribution of all the defect areas in the gray level image of the bottle bottom.
In the gray level image of the bottle bottom, the product of the shortest distance from the center of gravity of each defective area to the center of the circle of the outermost circular curve of the bottle bottom and the number of all pixel points in the defective area is recorded as the severity of each defective area;
Performing linear normalization processing on the severity of each defect area by using a norm () function to obtain the normalized severity of each defect area, wherein norm () is a known technique for normalizing the data value to be within the [0,1] interval;
Marking the average value of the normalized severity of all the defect areas as the defect degree of the detected bottle bottom;
And marking the bottle bottom of the outer package with the defect degree smaller than the preset defect area threshold as a qualified outer package, thereby completing the intelligent defect detection of the milk outer package.
The present invention has been completed.
To sum up, in the embodiment of the invention, the bottle bottom gray level image is obtained and the suspected region is obtained, the 8 neighborhood gray level uniformity of each pixel point of the bottle bottom gray level image is calculated, then the target region is screened out, a plurality of high-bright points are screened out in each target region, a plurality of defect regions are screened out in all target regions according to the 8 neighborhood gray level uniformity and gradient of the high-bright points and surrounding pixel points, the missed defect regions are searched according to the defect region characteristics, and whether the outer package of the bottle bottom is a qualified outer package is judged according to the position distribution of all defect regions in the bottle bottom gray level image. According to the method, the interference information is screened according to the gray level difference and the edge characteristics among the areas, the defect area is finally determined, the accuracy and the efficiency of identifying the defects of the bottled milk outer package in the prior art are effectively improved, and the loss of detail information is reduced.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the invention, but any modifications, equivalent substitutions, improvements, etc. within the principles of the present invention should be included in the scope of the present invention.
Claims (4)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411083204.6A CN119067923B (en) | 2024-08-08 | 2024-08-08 | Intelligent defect detection method for milk outer package |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411083204.6A CN119067923B (en) | 2024-08-08 | 2024-08-08 | Intelligent defect detection method for milk outer package |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119067923A CN119067923A (en) | 2024-12-03 |
| CN119067923B true CN119067923B (en) | 2025-02-07 |
Family
ID=93639683
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411083204.6A Active CN119067923B (en) | 2024-08-08 | 2024-08-08 | Intelligent defect detection method for milk outer package |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119067923B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541588A (en) * | 2024-01-10 | 2024-02-09 | 大连建峰印业有限公司 | Printing defect detection method for paper product |
| CN117745715A (en) * | 2024-02-06 | 2024-03-22 | 中科院南京耐尔思光电仪器有限公司 | Large-caliber telescope lens defect detection method based on artificial intelligence |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5772675B2 (en) * | 2012-03-23 | 2015-09-02 | トヨタ車体株式会社 | Gray image edge extraction method, edge extraction device, and gray image edge extraction program |
| CN112348773B (en) * | 2020-09-28 | 2022-09-23 | 歌尔股份有限公司 | Screen defect detection method and device and electronic equipment |
| CN115035122B (en) * | 2022-08-12 | 2022-11-11 | 宁波鑫芯微电子科技有限公司 | Artificial intelligence-based integrated circuit wafer surface defect detection method |
| CN116843678B (en) * | 2023-08-28 | 2023-11-21 | 青岛冠宝林活性炭有限公司 | Hard carbon electrode production quality detection method |
| CN117152129B (en) * | 2023-10-27 | 2024-02-13 | 智翼博智能科技(苏州)有限公司 | Visual detection method and system for surface defects of battery cover plate |
| CN117611563B (en) * | 2023-12-04 | 2024-05-10 | 东莞市隆亨钧精密制造有限公司 | Plastic mold production defect detection method based on image processing |
| CN118096736B (en) * | 2024-04-23 | 2024-06-25 | 山东德源电力科技股份有限公司 | Shell surface defect detection method for concentrator terminal production |
-
2024
- 2024-08-08 CN CN202411083204.6A patent/CN119067923B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541588A (en) * | 2024-01-10 | 2024-02-09 | 大连建峰印业有限公司 | Printing defect detection method for paper product |
| CN117745715A (en) * | 2024-02-06 | 2024-03-22 | 中科院南京耐尔思光电仪器有限公司 | Large-caliber telescope lens defect detection method based on artificial intelligence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119067923A (en) | 2024-12-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114359270B (en) | Computer vision-based automobile engine oil way copper sleeve defect detection method | |
| CN114937055B (en) | Image self-adaptive segmentation method and system based on artificial intelligence | |
| CN115082419B (en) | Blow-molded luggage production defect detection method | |
| CN116596936B (en) | Solar wing sailboard quality detection method based on image data | |
| CN114862849B (en) | Aluminum alloy plate film coating effect evaluation method based on image processing | |
| CN115082683A (en) | Injection molding defect detection method based on image processing | |
| CN113506286A (en) | Microwave chip defect detection method based on YOLOv5 algorithm small sample data set | |
| CN109685760B (en) | MATLAB-based SLM powder bed powder laying image convex hull depression defect detection method | |
| CN115147409B (en) | Mobile phone shell production quality detection method based on machine vision | |
| CN108827969A (en) | Metal parts surface defects detection and recognition methods and device | |
| CN114219805A (en) | Intelligent detection method for glass defects | |
| CN116523923B (en) | Battery case defect identification method | |
| CN116152242B (en) | Visual detection system of natural leather defect for basketball | |
| CN115953409B (en) | Injection molding surface defect detection method based on image processing | |
| CN110866915A (en) | Circular inkstone quality detection method based on metric learning | |
| CN114187289B (en) | Plastic product shrinkage pit detection method and system based on computer vision | |
| CN115330760B (en) | Diode pin defect identification method | |
| CN110060239B (en) | A kind of defect detection method for bottle mouth | |
| CN116630304B (en) | Lithium battery mold processing detection method and system based on artificial intelligence | |
| CN114708325A (en) | Method for quickly positioning rubber production problem based on rubber blooming defect | |
| CN115984272A (en) | Semitrailer axle defect identification method based on computer vision | |
| CN111968082A (en) | Product packaging defect detection and identification method based on machine vision | |
| CN117388263A (en) | Hardware terminal quality detection method for charging gun | |
| CN115131356A (en) | Steel plate defect classification method based on richness | |
| CN119067923B (en) | Intelligent defect detection method for milk outer package |
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 |