Injection molding part grinding quality detection method and system

Abstract

The invention relates to the technical field of image processing, in particular to an injection molding part grinding quality detection method and system. The method comprises the steps of collecting an injection molding part polishing image, obtaining a gradient magnitude of each pixel point in the injection molding part polishing image and surrounding pixel points of each pixel point, obtaining local feature intensity of each pixel point based on the gradient magnitude and the surrounding pixel points, and obtaining a local feature intensity of each pixel point according to the difference of the local feature intensity between the pixel points. Combining the pixel points to obtain a plurality of blocks, adaptively obtaining an optimal low threshold value of each block according to the local feature intensity of the pixel points in each block, and obtaining a high threshold value of each block according to the optimal low threshold value of each block; according to the optimal low threshold value and the optimal high threshold value of each block, edge pixel points of each block are obtained, according to the edge pixel points in each block, the grinding quality of the injection molding part is evaluated, and the accuracy of defect detection is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a method and system for detecting the grinding quality of injection-molded parts. Background Art

[0002] The detection of the grinding quality of injection-molded parts is an important link in ensuring product quality in the manufacturing industry. Traditional methods mainly rely on manual visual inspection, which is inefficient and limited by the experience and fatigue of operators, and is prone to misjudgment or missed inspection. With the development of image processing technology, automated detection systems have gradually been applied to the detection of the grinding quality of injection-molded parts, and defects are identified by analyzing the images of the surfaces of injection-molded parts.

[0003] During the detection process of the grinding quality of injection-molded parts, the Canny edge detection algorithm is usually used to detect the edges of defect areas. The Canny edge detection algorithm identifies edges by calculating the gradient magnitudes of each pixel point in the grinding image of the injection-molded part and by judging the relationship between the gradient magnitude of the pixel point and the preset high threshold and low threshold; it is known that the high threshold is the gradient magnitude threshold for determining strong edges, and pixel points with gradient magnitudes greater than the high threshold are marked as "strong edge pixel points", the low threshold is the gradient magnitude threshold for determining weak edges, and pixel points with gradient magnitudes between the high threshold and the low threshold are marked as "weak edge pixel points", and pixel points less than the low threshold are discarded, thereby identifying the defect edges in the grinding image of the injection-molded part; Since there are defect features with different intensities in the grinding image of the injection-molded part, that is, some defects have a large difference in gray value from the surrounding area, and some defects have a small difference in gray value from the surrounding area, the fixed high and low thresholds set by the traditional Canny edge detection algorithm cannot well identify the defect features. For defect areas with obvious gray value changes, the fixed threshold may be too high, resulting in these features being ignored; for defect areas with unobvious gray value changes, the fixed threshold may be too low, and some non-defect edges may be misidentified as defect features, thus leading to inaccurate quality detection results. Summary of the Invention

[0004] To solve the technical problem that the fixed high and low thresholds set by the Canny edge detection algorithm cannot well identify defects, the present invention provides a method and system for detecting the grinding quality of injection-molded parts.

[0005] In a first aspect, the present invention provides a method for detecting the grinding quality of injection-molded parts, adopting the following technical solution: A method for detecting the grinding quality of injection-molded parts includes the steps of: Collect the grinding image of the injection-molded part; Obtain the gradient magnitude of each pixel point in the polished image of the injection molded part and the surrounding pixel points of each pixel point, and obtain the local feature intensity of each pixel point according to the gradient magnitude and the surrounding pixel points; Merge the pixel points in the polished image of the injection molded part according to the difference in the local feature intensity between the pixel points to obtain a number of blocks; obtain the optimal low threshold of each block according to the mean value of the local feature intensity of all pixel points in each block; obtain the high threshold of each block according to the optimal low threshold of each block; obtain the edge pixel points of each block according to the optimal low threshold and the high threshold of each block; Evaluate the quality of the polished injection molded part according to the edge pixel points in each block.

[0006] The innovation of the present invention is that according to the local features of each pixel point, the local feature intensity of each pixel point is obtained, and then according to the difference in the local feature intensity between the pixel points during quality inspection, the pixel points are merged, so that defects of different degrees can be respectively merged into each block, which is convenient for adaptively obtaining the optimal low threshold of each block in the follow-up; then, according to the local feature intensity of the pixel points in each block, the optimal low threshold of each block is adaptively obtained. For the blocks with large defect degree, the low threshold is increased to avoid detecting details that do not belong to the defects. For the blocks with small defect degree, the low threshold is downgraded to facilitate detecting the weak defect edges and improve the accuracy of defect detection.

[0007] Preferably, the obtaining the gradient magnitude of each pixel point in the polished image of the injection molded part and the surrounding pixel points of each pixel point includes: Use the Sobel operator to obtain the gradient magnitude of each pixel point in the polished image of the injection molded part, and record the eight-neighborhood pixel points of each pixel point in the polished image of the injection molded part as the surrounding pixel points of each pixel point.

[0008] Preferably, the obtaining the local feature intensity of each pixel point includes: ; In the formula, represents the local feature intensity of the i-th pixel point; represents the gradient magnitude of the i-th pixel point; represents the number of surrounding pixel points of the i-th pixel point; represents the gradient magnitude of the j-th surrounding pixel point of the i-th pixel point; norm() represents the normalization function.

[0009] It can reflect the local features of each pixel point, which is convenient for subsequently merging pixel points with similar features to obtain blocks.

[0010] Preferably, the obtaining a number of blocks includes: Preset a feature threshold Q1. Denote the absolute value of the difference between the local feature intensity of the i-th pixel and that of each pixel within its four-neighbor range as the feature difference value between the i-th pixel and each of its four-neighbor pixels. Denote all four-neighbor pixels with a feature difference value less than Q1 as the target pixels of the i-th pixel. Merge the i-th pixel with all its target pixels to obtain the i-th block. Obtain the target pixels of each pixel in the i-th block and merge them with the i-th block to obtain a new i-th block. And so on, until there are no target pixels for each pixel in the new i-th block. Denote the new i-th block as the i-th block. Obtain each block.

[0011] It can merge the pixels belonging to the same region together to obtain each block, which is convenient for adaptively obtaining the optimal low threshold for each block subsequently.

[0012] Preferably, obtaining the optimal low threshold for each block includes: ; where, represents the optimal low threshold of the b-th block; represents the preset low threshold; represents the mean value of the local feature intensities of all pixels in the b-th block; represents the preset local feature intensity reference value; represents the preset hyperparameter.

[0013] For blocks with a large defect degree, increase the low threshold to avoid detecting details that do not belong to the defect. For blocks with a small defect degree, lower the low threshold to facilitate detecting weak defect edges and improve the accuracy of defect detection.

[0014] Preferably, obtaining the high threshold for each block according to the optimal low threshold of each block includes: Take twice the optimal low threshold of each block as the high threshold of each block.

[0015] Preferably, obtaining the edge pixels of each block according to the optimal low threshold and high threshold of each block includes: Take the pixels in the b-th block with a gradient magnitude greater than the high threshold as the strong edge pixels of the b-th block, take the pixels in the b-th block with a gradient magnitude between the high threshold and the low threshold as the weak edge pixels of the b-th block, and take the strong edge pixels of the b-th block and the weak edge points existing in the eight-neighbor range of its strong edge pixels as the edge pixels in the b-th block.

[0016] Preferably, evaluating the grinding quality of the injection molding part according to the edge pixels in each block includes: Preset a threshold T2 for the number of pixel points; based on the edge pixel points in each block, obtain several edges of the injection molded part grinding image. If the total number of pixel points of any edge of the injection molded part grinding image is greater than the threshold T2 for the number of pixel points, then mark this edge as the target edge of the injection molded part grinding image, and obtain all the target edges of the injection molded part grinding image; When the ratio of the number of pixel points of all the target edges of the injection molded part grinding image to the number of pixel points in the injection molded part grinding image is greater than 0.1, the system issues a warning to remind the staff that the quality of the injection molded part grinding is unqualified.

[0017] The accuracy of defect detection is improved, and defects of different degrees can be detected.

[0018] Preferably, the acquisition of the injection molded part grinding image includes: Use a high-resolution CMOS camera to align and shoot the surface of the injection molded part after grinding, collect the RGB image of the injection molded part grinding, and after graying the RGB image of the injection molded part grinding, record it as the injection molded part grinding image.

[0019] In a second aspect, the present invention provides an injection molded part grinding quality detection system, adopting the following technical solution: An injection molded part grinding quality detection system includes: a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned injection molded part grinding quality detection method is implemented.

[0020] By adopting the above technical solution, the above-mentioned injection molded part grinding quality detection method is generated into a computer program and stored in the memory to be loaded and executed by the processor, so as to manufacture a terminal device according to the memory and the processor, which is convenient to use.

[0021] The present invention has the following technical effects: The purpose of the present invention is to obtain the local feature intensity of each pixel point according to the local features of each pixel point, and then merge the pixel points according to the difference in the local feature intensity between the pixel points, so that defects of different degrees can be merged into each block respectively, which is convenient for adaptively obtaining the optimal low threshold of each block in the follow-up; then, according to the local feature intensity of the pixel points in each block, adaptively obtain the optimal low threshold of each block, increase the low threshold for the block with a large defect degree to avoid detecting details that do not belong to the defect, and lower the low threshold for the block with a small defect degree to facilitate detecting the weak defect edge, thereby improving the accuracy of defect detection. Description of the Drawings

[0022] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0023] Figure 1 The present invention is a method flow chart of a method for inspecting the polishing quality of injection molded parts according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments.

[0025] The embodiment of the present invention discloses a method for inspecting the polishing quality of injection molded parts. Figure 1 , comprising steps S1 to S4: S1: Collect images of injection molded parts being polished.

[0026] It should be noted that the present invention firstly needs to collect the surface image of the injection molded part after polishing, and then evaluate the polishing quality of the injection molded part by using an image processing algorithm.

[0027] In the implementation of the present invention, a high-resolution CMOS camera is used to shoot the polished surface of the injection molded part to collect the polished RGB image of the injection molded part. It should be noted that the focal length and aperture are adjusted during shooting to ensure clear focus on the surface of the injection molded part. The injection molded part polishing RGB image is grayed and recorded as the injection molded part polishing image.

[0028] S2: Obtain the gradient amplitude of each pixel point in the injection molded part polishing image and the surrounding pixels of each pixel point, and obtain the local feature intensity of each pixel point based on the gradient amplitude and the surrounding pixels of each pixel point.

[0029] It should be noted that in the process of detecting the grinding quality of injection molded parts, the Canny edge detection algorithm is usually used to detect the edges of the defect areas. The Canny edge detection algorithm calculates the gradient amplitude of each pixel point in the grinding image of the injection molded part, and identifies the edges by judging the relationship between the gradient amplitude of the pixel point and the preset high threshold and low threshold. Since there are defect features with different intensities in the grinding image of the injection molded part, that is, some defects have a large gray value difference from the surrounding area, and some defects have a small gray value difference from the surrounding area. Therefore, the fixed high and low thresholds set by the traditional Canny edge detection algorithm cannot well identify the defect features. For the defect areas with obvious gray value changes, the fixed threshold is too high, resulting in these features being ignored; for the defect areas with unobvious gray value changes, the fixed threshold is too low, and some non-defect edges will be misidentified as defect features, resulting in inaccurate quality detection results; Therefore, the object of the present invention is to divide the normal area (the area with better grinding quality of the injection molded part), the defect area with obvious gray value change, and the defect area with unobvious gray value change in the grinding image of the injection molded part into a separate block respectively, and further perform edge detection by adaptively setting the low threshold for each block according to the characteristics of each block; Since the gray value distribution of the pixel points in the normal area of the grinding image of the injection molded part is relatively uniform, the gradient amplitude of the pixel points in the normal area of the grinding image of the injection molded part is relatively low, and the gradient value difference between the pixel points in the normal area and their surrounding pixel points is relatively small; while the gradient values of the pixel points in the defect area with obvious gray value change and the defect area with unobvious gray value change in the grinding image of the injection molded part are larger than those in the normal area, and there is a large difference between the gradient values of the pixel points and their surrounding pixel points. Therefore, by analyzing the gradient amplitude of each pixel point and the gradient amplitude difference between each pixel point and its surrounding pixel points, the local feature intensity of each pixel point is obtained.

[0030] It should be further noted that since the gray value change degrees of different defect areas are different, the local feature intensities of the pixel points in the defect areas with different gray value changes are different, and the local feature intensities of the pixel points in the same defect area are approximate. Therefore, when dividing blocks according to the local feature intensity of each pixel point subsequently, the normal area, the defect area with obvious gray value change, and the defect area with unobvious gray value change can be divided into a separate block respectively.

[0031] In the embodiment of the present invention, the Sobel operator is used to obtain the gradient amplitude of each pixel point in the grinding image of the injection molded part, and the eight-neighborhood pixel points of each pixel point in the grinding image of the injection molded part are denoted as the surrounding pixel points of each pixel point; Obtain the local feature intensity of each pixel point: ; In the formula, represents the local feature intensity of the i-th pixel point; represents the gradient magnitude of the i-th pixel point; represents the number of surrounding pixel points of the i-th pixel point; represents the gradient magnitude of the j-th surrounding pixel point of the i-th pixel point; norm() represents the normalization function; represents the gradient magnitude of the pixel point. The larger its value, the greater the local feature intensity of the pixel point; represents the difference in gradient magnitude between the i-th pixel point and its surrounding pixel points. The smaller its value, the greater the difference in gradient magnitude between the i-th pixel point and its surrounding pixel points, indicating that the gradient change between the i-th pixel point and its surrounding pixel points is relatively large, and thus the local feature intensity of the i-th pixel point is greater.

[0032] It should be noted that for the area with good grinding quality of the injection molded part, the gray value distribution of the pixel points in this area is relatively uniform. Therefore, if the gradient magnitude of any pixel point is small and the difference in gradient magnitude between this pixel point and its neighboring pixel points is small, the local feature intensity of this pixel point is small, and this pixel point is more likely to belong to the area with good grinding quality of the injection molded part, that is, the normal area.

[0033] For the area with poor grinding quality of the injection molded part, that is, the area with defects such as scratches, the gray value distribution of the pixel points in this area and their neighboring pixel points is not uniform. Therefore, if the gradient magnitude of any pixel point is large and the difference in gradient magnitude between this pixel point and its neighboring pixel points is large, the local feature intensity of this pixel point is greater, and this pixel point is more likely to belong to the area with poor grinding quality of the injection molded part, that is, the defective area.

[0034] S3: According to the difference in local feature intensity between pixel points, merge the pixel points in the grinding image of the injection molded part to obtain several blocks. According to the local feature intensity of the pixel points in each block, adaptively obtain the optimal low threshold of each block. According to the optimal low threshold of each block, obtain the high threshold of each block; according to the optimal low threshold and high threshold of each block, obtain the edge pixel points of each block.

[0035] It should be noted that the above steps obtain the local feature intensity of each pixel point. Then, according to the difference in local feature intensity between pixel points before, merge the pixel points to obtain several blocks in the grinding image of the injection molded part. If the difference in local feature intensity between pixel points is small, it means that the pixel points should be merged.

[0036] In an embodiment of the present invention, a feature threshold Q1 = 0.3 is preset. The absolute value of the difference between the local feature intensity of the i-th pixel and each pixel within its four-neighborhood range is denoted as the feature difference value between the i-th pixel and each of its four-neighborhood pixels. All four-neighborhood pixels with the feature difference value less than Q1 are denoted as the target pixels of the i-th pixel; the i-th pixel is merged with all its target pixels to obtain the i-th sub-block; the target pixels of each pixel in the i-th sub-block are obtained and merged with the i-th sub-block to obtain a new i-th sub-block; and so on, until there are no target pixels for each pixel in the new i-th sub-block, and the new i-th sub-block is denoted as the i-th block. Any pixel other than the pixels of the i-th block in the polished injection molding part image is denoted as the n-th pixel, and the i-th pixel is expanded to obtain the n-th block; and so on, until there are no pixels in the polished injection molding part image, and each block is obtained.

[0037] It should be noted that the local feature intensity of pixels in a defect area with obvious gray-scale changes is relatively large. Therefore, if the pixels in any sub-block are more likely to be in a defect area with obvious gray-scale changes, the low threshold needs to be appropriately increased at this time to avoid detecting some details of non-defect edges in the defect area. The local feature intensity of pixels in a defect area with unobvious gray-scale changes is smaller than that of pixels in a defect area with obvious gray-scale changes. Therefore, if the pixels in any sub-block are more likely to be in a defect area with unobvious gray-scale changes, the low threshold needs to be appropriately decreased at this time to detect some details of weak defect edges in the defect area. Therefore, in the present invention, a local feature intensity reference value is preset. If the mean value of the local feature intensity of pixels in any block is greater than the local feature intensity reference value, it is considered that this block is more likely to be a defect area with obvious gray-scale changes, and the low threshold of this block is increased at this time; if the mean value of the local feature intensity of pixels in any block is less than the local feature intensity reference value, it is considered that this block is more likely to be a defect area with unobvious gray-scale changes or a normal area, and the low threshold of this block is decreased at this time. When this block is a defect area with unobvious gray-scale changes, decreasing the low threshold can detect weak edges, and when this block is a normal area with gray-scale distribution, decreasing the low threshold cannot detect edges either.

[0038] In an embodiment of the present invention, a low threshold is preset , and in other embodiments, the implementer can preset the value of the low threshold according to the specific implementation.

[0039] Obtain the optimal low threshold of each block: ; In the formula, represents the optimal low threshold of the b-th sub-block; represents a preset low threshold; represents the mean of the local feature intensities of all pixel points in the b-th block; represents a preset local feature intensity reference value. In the embodiments of the present invention, the preset local feature intensity reference value , and in other embodiments, the implementer can preset according to the specific implementation situation; represents a preset hyperparameter. In the embodiments of the present invention, the preset , and in other embodiments, the implementer can preset according to the specific implementation situation; known The value of represents the adjustment amount of the preset low threshold; if the value of is greater than , it indicates that the mean of the local feature intensities of all pixel points in the b-th block is relatively large. At this time, the pixel points in the b-th block are more likely to be defect regions with obvious gray-scale changes. At this time, it is necessary to appropriately increase the low threshold to avoid detecting some details of non-defect edges in the defect region; If the value of is less than , it indicates that the mean of the local feature intensities of all pixel points in the b-th block is relatively small. At this time, the pixel points in the b-th block are more likely to be defect regions with unobvious gray-scale changes or normal regions. At this time, it is necessary to appropriately reduce the low threshold to detect some details of weak edges in the unobvious defect region. The gray-scale value distribution of the normal region is uniform, and reducing the low threshold has no impact.

[0040] The function of is to adjust the adjustment amount of the preset low threshold to avoid the adjustment amount of the preset low threshold being too large or too small, resulting in the preset low threshold being adjusted too large or too small.

[0041] It should be noted that according to the low threshold of each block, the high threshold of each block is obtained. Therefore, based on the low threshold and high threshold of each block, edge detection is performed on each block to obtain the edge pixel points of each block, that is, several edges of the injection molded part grinding image are obtained.

[0042] In the embodiments of the present invention, twice the optimal low threshold of the b-th block is used as the high threshold of the b-th block. The pixel points in the b-th block with a gradient amplitude greater than the high threshold are used as the strong edge pixel points of the b-th block. The pixel points in the b-th block with a gradient amplitude between the high threshold and the low threshold are used as the weak edge pixel points of the b-th block. The strong edge pixel points of the b-th block and the weak edge points existing in the eight-neighborhood range of its strong edge pixel points are used as the edge pixel points in the b-th block.

[0043] Similarly, obtain the edge pixel points in each block.

[0044] S4: Evaluate the polishing quality of the injection molded part according to the edge pixel points in each block.

[0045] It should be noted that first, based on the edge pixel points in each block, several edges of the injection molded part polishing image are obtained, and the noise edges in the injection molded part polishing image are screened to obtain the target edges in the injection molded part polishing image, that is, the defect edges; if the number of pixel points on the target edges in the injection molded part polishing image is larger, it indicates that the polishing quality of the injection molded part is worse.

[0046] In the embodiment of the present invention, the preset pixel point number threshold T2 = 10. In other embodiments, the implementer can preset the value of T2 according to the specific implementation situation; based on the edge pixel points in each block, several edges of the injection molded part polishing image are obtained. If the total number of pixel points on any edge of the injection molded part polishing image is greater than the pixel point number threshold T2, then mark this edge as the target edge of the injection molded part polishing image, and obtain all the target edges of the injection molded part polishing image; If the ratio of the number of pixel points of all target edges to the number of pixel points in the injection molded part polishing image is greater than 0.1, the system issues an alarm to remind the staff that the polishing quality of the injection molded part is unqualified.

[0047] The embodiment of the present invention also discloses an injection molded part polishing quality detection system, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, it implements an injection molded part polishing quality detection method according to the present invention.

Claims

1. A method for detecting the grinding quality of injection molded parts, characterized in that, Including the steps: Collect the grinding image of the injection molded part; Obtain the gradient amplitude of each pixel point in the grinding image of the injection molded part and the surrounding pixel points of each pixel point, and obtain the local feature intensity of each pixel point according to the gradient amplitude and the surrounding pixel points; Merge the pixel points in the grinding image of the injection molded part according to the difference in the local feature intensity between the pixel points to obtain several blocks; obtain the optimal low threshold of each block according to the mean value of the local feature intensity of all pixel points in each block; obtain the high threshold of each block according to the optimal low threshold of each block; obtain the edge pixel points of each block according to the optimal low threshold and the high threshold of each block; Evaluate the grinding quality of the injection molded part according to the edge pixel points in each block.

2. The method for detecting the grinding quality of an injection molded part according to claim 1, characterized in that, The obtaining the gradient amplitude of each pixel point in the grinding image of the injection molded part and the surrounding pixel points of each pixel point includes: Use the Sobel operator to obtain the gradient amplitude of each pixel point in the grinding image of the injection molded part, and record the eight-neighborhood pixel points of each pixel point in the grinding image of the injection molded part as the surrounding pixel points of each pixel point.

3. The injection molded part grinding quality inspection method according to claim 1, characterized in that, The obtaining the local feature intensity of each pixel point includes: ； In the formula, represents the local feature intensity of the i-th pixel point; represents the gradient amplitude of the i-th pixel point; represents the number of surrounding pixel points of the i-th pixel point; represents the gradient amplitude of the j-th surrounding pixel point of the i-th pixel point; norm() represents the normalization function.

4. A method for detecting the grinding quality of an injection molded part according to claim 1, characterized in that, The obtaining several blocks includes: Preset a feature threshold Q1, record the absolute value of the difference in the local feature intensity between the i-th pixel point and each pixel point within its four-neighborhood range as the feature difference value between the i-th pixel point and each of its four-neighborhood pixel points, and record all four-neighborhood pixel points with the feature difference value less than Q1 as the target pixel points of the i-th pixel point; merge the i-th pixel point with all its target pixel points to obtain the i-th sub-block; obtain the target pixel points of each pixel point in the i-th sub-block and merge them with the i-th sub-block to obtain a new i-th sub-block; and so on, until there are no target pixel points for each pixel point in the new i-th sub-block, record the new i-th sub-block as the i-th block; obtain each block.

5. A method for detecting the grinding quality of an injection molded part according to claim 1, characterized in that, The obtaining the optimal low threshold of each block includes: ； In the formula, represents the optimal low threshold of the b-th block; represents the preset low threshold; represents the mean of the local feature intensities of all pixel points in the b-th block; represents the preset local feature intensity reference value; represents the preset hyperparameter.

6. The method for detecting the grinding quality of an injection molded part according to claim 1, characterized in that, The obtaining the high threshold of each block according to the optimal low threshold of each block includes: Take twice the optimal low threshold of each block as the high threshold of each block.

7. A method for detecting the grinding quality of an injection molded part according to claim 1, characterized in that, The obtaining the edge pixel points of each block according to the optimal low threshold and the high threshold of each block includes: Take the pixel points with a gradient amplitude greater than the high threshold in the b-th block as the strong edge pixel points of the b-th block, take the pixel points with a gradient amplitude between the high threshold and the low threshold in the b-th block as the weak edge pixel points of the b-th block, and take the strong edge pixel points of the b-th block and the weak edge points existing in the eight-neighborhood range of its strong edge pixel points as the edge pixel points in the b-th block.

8. A method for detecting the grinding quality of an injection molded part according to claim 1, characterized in that, The evaluating the grinding quality of the injection molded part according to the edge pixel points in each block includes: Preset a pixel point number threshold T2; obtain several edges of the grinding image of the injection molded part according to the edge pixel points in each block, if the total number of pixel points of any edge of the grinding image of the injection molded part is greater than the pixel point number threshold T2, record this edge as the target edge of the grinding image of the injection molded part, and obtain all the target edges of the grinding image of the injection molded part; When the ratio of the number of pixel points of all target edges in the injection molded part grinding image to the number of pixel points in the injection molded part grinding image is greater than 0.1, the system issues a warning to remind the staff that the quality of the injection molded part grinding is unqualified.

9. A method for detecting the grinding quality of an injection molded part according to claim 1, characterized in that, The acquisition of the injection molded part grinding image includes: Using a high-resolution CMOS camera to align and shoot the surface of the injection molded part after grinding, collecting the injection molded part grinding RGB image, and after graying the injection molded part grinding RGB image, it is recorded as the injection molded part grinding image.

10. An injection molded part grinding quality detection system, characterized in that, Including: A processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for detecting the quality of injection molded part grinding according to any one of claims 1-9 is implemented.

Images