Injection molding panel production quality detection method and system based on image enhancement

Abstract

The invention relates to the technical field of image processing, in particular to an injection molding panel production quality detection method and system based on image enhancement. The method comprises the following steps: calculating the throwing degree of each pixel point; taking the product of the throwing degree and the standard deviation preset in the bilateral filter as the optimal standard deviation of each pixel point; processing and detecting the image based on the optimal standard deviation; the step of calculating the throwing degree of each pixel point comprises the following steps of: calculating the neighborhood gray scale confusion degree of each pixel point; obtaining a point cluster containing edge information in the image, and calculating gradient consistency between pixel points in the point cluster and other pixel points in the point cluster; setting a weight coefficient of gradient consistency, and determining the regularity of the pixel point by taking the weighted sum of a plurality of gradient consistency as a total trend; and taking the ratio of the gradient law degree corresponding to the pixel point to the neighborhood gray scale confusion degree as the throwing degree of the pixel point. The method and the device have the effect of improving the accuracy of subsequent product anomaly detection.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method and system for detecting the production quality of injection-molded panels based on image enhancement. Background Art

[0002] An injection-molded panel refers to a plastic shell made by an injection molding process, which is mainly the outer shell of some electronic products. The surface quality of the injection-molded panel is directly related to the overall aesthetics of the product, so the surface quality of the injection-molded panel needs to be inspected before assembly. The traditional inspection method is manual observation by staff, but manual inspection has many defects, such as low efficiency, high missed detection rate, and heavy labor burden. Currently, with the development and progress of electronic information technology, the method of automatic inspection of injection-molded panels through machine vision has gradually replaced traditional manual inspection. In the process of inspecting injection-molded panels through machine vision, due to the complexity of the industrial cycle, after collecting the image of the injection-molded panel, the image needs to be denoised before subsequent inspection can be carried out.

[0003] Bilateral filtering is a nonlinear smoothing filtering technology commonly used in the field of image processing. It considers both spatial information and grayscale similarity, and can preserve edges and remove noise. During the bilateral filtering process, the standard deviation of the grayscale weight needs to be set. During the image processing process, a higher standard deviation of the grayscale weight will cause abnormal areas in the image to be over-smoothed, thereby increasing the difficulty of subsequent abnormal detection of injection-molded panels. A lower standard deviation will make it difficult to filter noise in the image. Therefore, in the process of bilateral filtering methods in related technologies for image processing of injection-molded panels with complex features, there is a situation where abnormal areas are over-smoothed or noise is not smoothed enough, which ultimately leads to lower accuracy in subsequent abnormal detection of injection-molded panels. Summary of the Invention

[0004] In order to reduce the low accuracy of abnormality detection caused by over-smoothing or insufficient smoothing during injection molding image processing, the present application provides an injection molding panel production quality detection method and system based on image enhancement.

[0005] In a first aspect, the present application provides a method for inspecting the production quality of injection molded panels based on image enhancement, which adopts the following technical solutions: A method for detecting the production quality of injection-molded panels based on image enhancement, comprising the steps of: calculating the reproducibility of each pixel; taking the product of the reproducibility and a preset standard deviation in a bilateral filter as the optimal standard deviation of each pixel; and processing and detecting the image based on the optimal standard deviation. The steps of calculating the degree of scalability of each pixel point include: calculating the grayscale chaos degree of the neighborhood of each pixel point; obtaining a point cluster containing edge information in the image, and calculating the gradient consistency between the pixel point in the point cluster and other pixel points in the point cluster; setting a weight coefficient of gradient consistency, and determining the regularity of the pixel point with the weighted sum of multiple gradient consistencies as the overall trend; taking the ratio of the gradient regularity corresponding to the pixel point to the grayscale chaos degree of the neighborhood as the scalability of the pixel point; wherein, the steps of calculating the gradient consistency of the pixel point include: obtaining the gradient direction of each pixel point in the point cluster, and for each pixel point, obtaining the gradient direction difference and gradient value difference between the pixel point and other pixel points; obtaining gradient consistency based on the gradient direction difference and gradient value difference, and the gradient direction difference and gradient value difference are both inversely proportional to the gradient consistency.

[0006] The beneficial effects are as follows: the grayscale disorder of the neighborhood of each pixel in the image is calculated. Based on the grayscale disorder of the pixels in the pixel's neighborhood, the amount of information near the pixel can be determined. Pixels with high information content should be retained to reduce the subsequent smoothing of abnormal areas. Furthermore, the present application distinguishes edges in the image. Pixels on the normal edges of the product in the image have a high degree of regularity. Excessive retention of these edges can affect the subsequent detection of abnormal edges. Therefore, pixels with high regularity are smoothed more frequently in the present application. The pixel's reproducibility is calculated by combining the neighborhood disorder and regularity. When the image is subsequently processed using bilateral filtering, the standard deviation of the grayscale weights in the bilateral filter is adjusted based on the reproducibility. This allows the standard deviation of the grayscale weights to be dynamically adjusted throughout the image processing process. Compared to the fixed, preset standard deviation of grayscale weights used in related art, this method can reduce the problem of undersmoothing in non-abnormal areas and oversmoothing in abnormal areas, ultimately improving the accuracy of injection molded product inspection.

[0007] Optionally, the steps for calculating the neighborhood grayscale chaos of a pixel point include: calculating the center difference of the pixel point based on the difference in grayscale values between the pixel point and the pixel points in its corresponding neighborhood; calculating the neighborhood difference of the pixel point based on the difference in grayscale values of each pixel point in the neighborhood; and taking the product of the neighborhood difference and the center difference as the neighborhood grayscale chaos of the pixel point.

[0008] The beneficial effect is that in the process of calculating the neighborhood chaos degree, on the one hand, the grayscale value difference between the pixel point and the pixel points in its corresponding neighborhood is compared, and on the other hand, the difference between each pixel point in the pixel point neighborhood is compared, which takes into account more comprehensiveness and makes the calculation of the neighborhood chaos degree more accurate.

[0009] Optionally, the central difference is calculated as: Where, For the image OK The difference in the center of the pixels in the column; is the number of pixels in the neighborhood; Indicates the Rank Grayscale value of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The grayscale value extremes of the pixel in the column and all the pixels in its neighborhood.

[0010] The beneficial effects are: Indicates the Rank The difference in grayscale value between a pixel and any pixel in its neighborhood. This difference is compared with the range of the grayscale values of the pixel itself and all pixels in its neighborhood to normalize the difference for ease of subsequent calculations. Finally, the differences between the pixel and all pixels in its corresponding neighborhood are accumulated and averaged to obtain the center difference of the pixel.

[0011] Optionally, the formula for calculating neighborhood differences is: Where, Indicates the Rank Neighborhood differences of column pixels; Indicates the Rank The grayscale value extremes of all pixels in the neighborhood of the column pixel; For the Rank The number of pixels in the neighborhood of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The neighborhood of the pixel in the column The grayscale value of a pixel.

[0012] The beneficial effects are: This part represents the difference in grayscale value between any two adjacent pixels in a pixel's neighborhood. By multiplying this difference, the final calculation result can reflect the fluctuation of the grayscale value of the pixels in the neighborhood, and thus be used to calculate the neighborhood disorder.

[0013] Optionally, the step of obtaining a point cluster containing edge information in the image includes: obtaining a binary image of the image using the Otsu threshold method; and taking each white area in the binary image as a point cluster.

[0014] The beneficial effect is that the image is processed using the Otsu thresholding method to obtain a binary image. The binary image consists of white and black areas. The white areas correspond to the original edges of the product in the image, abnormal scratches, abnormal cracks, and other edges. All pixels in the white areas are used as a set to form a point cluster containing edge information.

[0015] Optionally, the pixel regularity is calculated by normalizing the overall trend.

[0016] The beneficial effect is: normalizing the value of the overall trend to facilitate subsequent calculation and processing of the data.

[0017] The calculation formula of regularity is: Where, No. In the point cluster The regularity of pixels; For the The number of pixels in a cluster; Indicates the In the point cluster The overall trend of pixels.

[0018] Optionally, the calculation formula for the weight coefficient of the gradient consistency of the pixel points in the point cluster is: Where, Indicates the In the point cluster and The weight coefficient of the gradient consistency between pixels; Indicates the The first point in the cluster and The Euclidean distance between pixels.

[0019] The beneficial effect is that in the process of calculating the regularity of pixel points in a point cluster, the gradient information of other pixel points that are closer to the pixel point itself should be considered. Therefore, the weight coefficient is set for the pixel point and the other different pixel points in the point cluster according to the Euclidean distance between the two pixel points, thereby further improving the accuracy of subsequent regularity calculation.

[0020] Optionally, the step of performing abnormality detection on the image based on the optimal image includes: setting an abnormal length; detecting an edge of the optimal image, and determining that the product is defective when the edge length is greater than the abnormal length.

[0021] In a second aspect, the present application provides an injection molding panel production quality inspection system based on image enhancement, which adopts the following technical solutions: A system for inspecting the production quality of injection-molded panels based on image enhancement includes: a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the system implements the above-mentioned method for inspecting the production quality of injection-molded panels based on image enhancement.

[0022] The beneficial effect is that the above-mentioned injection molding panel production quality inspection method based on image enhancement is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor convenient to use.

[0023] This application has the following technical effects: In this application, pixels with more information are screened based on the neighborhood chaos corresponding to the pixel points, and the original edges of the product in the image and the abnormal edges are distinguished based on the regularity of the image edge pixels. The pixel's reproducibility is calculated by combining the neighborhood chaos and regularity. During the image processing process using a bilateral filter, the standard deviation of its grayscale weight is dynamically adjusted according to the reproducibility of each pixel point, thereby reducing the situation where abnormal areas in the image are oversmoothed or the non-abnormal areas have too much noise, effectively enhancing the characteristics of the abnormal areas in the image and improving the accuracy of subsequent abnormality detection of the product. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a method flow chart of a method for inspecting the production quality of injection molded panels based on image enhancement according to an embodiment of the present application.

[0025] Figure 2 This is a method flow chart of step S1 of a method for inspecting the production quality of injection-molded panels based on image enhancement in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The present invention discloses a method for inspecting the production quality of injection-molded panels based on image enhancement. The method calculates the smoothness of each pixel in an image and adjusts the standard deviation of grayscale weights during bilateral filtering based on the smoothness. During image processing using a bilateral filter, the standard deviation of grayscale weights is dynamically adjusted based on the smoothness of different pixels. This reduces the chances of over-smoothing or under-smoothing abnormal areas in the image, effectively enhancing the image's smoothness and improving the accuracy of anomaly detection in subsequent products.

[0027] Reference Figure 1 The method for detecting the production quality of injection molded panels based on image enhancement includes steps S1 and S2.

[0028] S1: Calculate the reproducibility of each pixel, and take the product of the reproducibility and the preset standard deviation in the bilateral filter as the optimal standard deviation of each pixel; When using a bilateral filter to process an image, a smaller grayscale weight standard deviation should be used for processing abnormal areas with low sensitivity (such as bubbles, scratches, and cracks). For non-abnormal areas, a larger grayscale weight standard deviation can be used for smoothing to highlight the abnormal areas. This reduces interference from subsequent non-abnormal areas or noise on abnormality detection, ultimately improving the accuracy of product abnormality detection.

[0029] Reference Figure 2 , in one embodiment, step S1 includes: S11-step S14; S11: Calculate the grayscale disorder degree of the neighborhood of each pixel; Calculate the center difference of a pixel based on the difference in grayscale value between the pixel and the pixel in its corresponding neighborhood; The default size is The neighborhood of , in this embodiment, the size of the neighborhood window is , for each pixel in the image, its neighborhood includes 8 pixels; the center difference between each pixel and its neighboring pixels is calculated based on the grayscale value of the pixel itself and the pixels in its corresponding neighborhood; In one embodiment, the center difference is calculated as follows: Where, For the image OK The difference in the center of the pixels in the column; is the number of pixels in the neighborhood; Indicates the Rank Grayscale value of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The grayscale value extremes of the pixel in the column and all the pixels in its neighborhood.

[0030] Indicates the image Rank The grayscale value of the pixel in the column is the same as the grayscale value of the pixel in the corresponding neighborhood. The difference between the grayscale values of pixels; Calculated the Rank The sum of the grayscale differences between a column pixel and all pixels in its corresponding neighborhood. The larger the Rank The more inconsistent the grayscale of the pixel in the column is with the pixels in its neighborhood, the Rank The grayscale value changes of pixels in the neighborhood of the column pixel are more chaotic. Used for The value of is normalized to facilitate the subsequent calculation of the neighborhood grayscale chaos degree.

[0031] In another embodiment, the calculation formula of the center difference is: Where, For the image OK The difference in the center of the pixels in the column; is the number of pixels in the neighborhood; Indicates the Rank Grayscale value of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The grayscale value extremes of the pixel in the column and all the pixels in its neighborhood.

[0032] In order to adapt to the application environment with high requirements for calculation efficiency, absolute value operation is used in this formula to simplify the calculation steps and improve the efficiency of center difference calculation.

[0033] The neighborhood difference of the pixel point is calculated according to the gray value difference of each pixel point in the neighborhood, and the product of the neighborhood difference and the center difference is the neighborhood grayscale disorder of the pixel point.

[0034] In one embodiment, the neighborhood difference is calculated as follows: Where, Indicates the Rank Neighborhood differences of column pixels; Indicates the Rank The grayscale value extremes of the pixel in the column and all pixels in its neighborhood; For the Rank The number of pixels in the neighborhood of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The neighborhood of the pixel in the column The grayscale value of a pixel.

[0035] Indicates the difference in grayscale values between two adjacent pixels in the neighborhood; if the grayscale values of two adjacent pixels change smoothly, then If the grayscale values of two adjacent pixels change dramatically, then The value of is large, and then through The value can reflect the change in the grayscale of pixels in the neighborhood.

[0036] The difference between any two adjacent pixels in the neighborhood is multiplied to reflect the overall grayscale fluctuation of all pixels in the neighborhood. At the same time, the neighborhood difference is calculated by the multiplication method, which makes the final calculation result more sensitive to the changes in the grayscale value within the neighborhood.

[0037] In another embodiment, the neighborhood difference is calculated as follows: Where, Indicates the Rank Neighborhood differences of column pixels; Indicates the Rank The grayscale value extremes of the pixel in the column and all pixels in its neighborhood; For the Rank The number of pixels in the neighborhood of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The neighborhood of the pixel in the column The grayscale value of a pixel.

[0038] This formula uses the cumulative average method to calculate the neighborhood differences, making the output more stable and fast, and is suitable for scenarios where real-time and fast calculation of neighborhood differences is required.

[0039] After the neighborhood difference is calculated, the product of the neighborhood difference and the center difference is calculated to obtain the first Rank The grayscale chaos of the neighborhood corresponding to the pixel is calculated by combining neighborhood differences and center differences. This allows for a more comprehensive and accurate calculation of the grayscale chaos. A higher grayscale chaos indicates more information surrounding the pixel. Therefore, more grayscale detail should be retained during subsequent processing to reduce oversmoothing of abnormal areas and improve the accuracy of anomaly detection.

[0040] S12: Obtain a point cluster containing edge information in the image, and calculate the gradient consistency between a pixel in the point cluster and other pixels in the point cluster; In product images, abnormal areas are characterized by chaotic grayscale values and chaotic, complex, and irregular pixel gradients. Therefore, this application calculates the gradient consistency of pixel points to further screen and enhance abnormal areas, improving the accuracy of subsequent anomaly detection.

[0041] Obtain the gradient value and gradient direction of each pixel in the image, and use the Otsu thresholding method to perform threshold segmentation on the image to form a binary image. The binary image includes multiple white areas corresponding to product edges and defect edges, so that the multiple pixels in each white area form a point cluster.

[0042] Calculate the gradient consistency of each pixel in the point cluster. If the gradient consistency in the point cluster is poor, it means that the gradient value and gradient direction in the point cluster are chaotic, and the possibility of belonging to an abnormal area is high, so that the abnormal area can be further screened.

[0043] In one embodiment, the steps for calculating the gradient consistency of pixel points include: obtaining the gradient direction of each pixel point in the point cluster, and for each pixel point, obtaining the gradient direction difference and gradient value difference between the pixel point and other pixel points; obtaining gradient consistency based on the gradient direction difference and gradient value difference, and the gradient direction difference and gradient value difference are both inversely proportional to the gradient consistency.

[0044] Specifically, the calculation formula for the gradient consistency of pixels can be expressed as: ; Where, Indicates the The first point in the cluster pixels and the The gradient consistency between pixels, For the The gradient direction of each pixel; For the In the point cluster The gradient direction of each pixel, For the In the point cluster The gradient value of each pixel; For the In the point cluster The gradient value of a pixel.

[0045] Indicates the point cluster and The smaller the value of this part, the more consistent the gradient direction between the two pixels, which will lead to greater gradient consistency. make The value of is negatively correlated with the final gradient consistency, i.e. The larger it is, the less consistent the calculated gradient is.

[0046] It represents the difference in gradient values between two pixels. The gradient value indicates how fast the grayscale of a pixel changes. In the image, the gradient value of the original edge of the injection-molded panel is relatively consistent. However, the gradient value of the pixel in the abnormal area is unstable. The gradient values of each pixel are chaotic and fluctuate greatly. Therefore, this characteristic can be used to filter out abnormal areas. In this formula, Negatively correlated with gradient consistency, i.e. The larger the value, the less consistent the final gradient will be.

[0047] S13: Setting a weight coefficient for gradient consistency, and determining the regularity of the pixel point using the weighted sum of multiple gradient consistencies as the overall trend; In one embodiment, the calculation formula of the weight coefficient of the gradient consistency of the pixel points in the point cluster is: Where, Indicates the In the point cluster and The weight coefficient of the gradient consistency between pixels; Indicates the The first point in the cluster and The Euclidean distance between pixels.

[0048] In a point cluster, when the Euclidean distance between two pixels is close, more consideration should be given to the difference in gradient values between the two pixels and the consistency of gradient directions between the two pixels to improve the accuracy of the final calculation of regularity.

[0049] In another embodiment, the calculation formula of the weight coefficient of the gradient consistency of the pixel points in the point cluster is: Where, Indicates the In the point cluster and The weight coefficient of the gradient consistency between pixels; Indicates the The first point in the cluster and The Euclidean distance between pixels; is the adjustment coefficient.

[0050] The adjustment coefficient is introduced in this formula. On the one hand, it is convenient for the staff to adjust the weight according to the actual situation during the work process. At the same time, after setting the adjustment coefficient, the final value of the weight coefficient can decrease exponentially with the increase of the distance between the two pixel points, thereby increasing the sensitivity of the weight coefficient to the distance between the two pixel points.

[0051] The gradient consistency is weighted and then accumulated to obtain the overall trend of the pixels in the cluster.

[0052] Get the The gradient consistency between the pixel point and the pixel points in the cluster except itself is weighted according to their respective corresponding weight coefficients, and then accumulated to obtain the first In the point cluster Similarly, the overall trend of the remaining pixels in the point cluster is obtained respectively.

[0053] Specifically, the calculation formula of the overall trend can be expressed as: Where, Indicates the In the point cluster The overall trend of pixels Indicates the In the point cluster and The weight coefficient of the gradient consistency between pixels; Indicates the The first point in the cluster pixels and the Gradient consistency between pixels; Indicates the The number of pixels in a cluster.

[0054] The value normalized by the overall trend is the regularity of the pixel points; Specifically, the calculation formula for regularity is: Where, For the In the point cluster The regularity of pixels; For the The number of pixels in a cluster; .

[0055] The overall trend is normalized to facilitate subsequent calculations.

[0056] S14: The ratio of the gradient regularity of the pixel point to the grayscale disorder of the neighborhood is the throwability of the pixel point; There is no regularity in the pixels in the black area of the binary image, so the regularity of the pixels in the black area of the binary image is fixed to a natural constant. In this embodiment, the regularity of the pixels in the black area of the binary image is fixed to 1.

[0057] S2: Process and detect the image based on the optimal standard deviation.

[0058] Set an abnormal length; detect the edge of the optimal image, and determine that the product is defective when the edge length is greater than the abnormal length. In this embodiment, the abnormal length is 10. In other embodiments, the abnormal length can be adjusted according to the actual product and the type of defect that is likely to occur.

[0059] After processing the product image according to the degree of distortion, the noise and the original edge pixels of the product in the image are more smoothly faded, so that the edge pixels in the abnormal area are more detailed, making the abnormality detection of the image more accurate.

[0060] An embodiment of the present application also discloses an injection molding panel production quality inspection system based on image enhancement, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, an injection molding panel production quality inspection method based on image enhancement according to the present application is implemented.

[0061] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0062] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for detecting the production quality of injection molded panels based on image enhancement, characterized in that: include: Step S1: Calculate the reproducibility of each pixel; The product of the reproducibility and the preset standard deviation in the bilateral filter is taken as the optimal standard deviation of each pixel; Step S2: Process and detect the image based on the optimal standard deviation; Step S1 includes: step S11: calculating the neighborhood grayscale chaos of each pixel point; step S12: obtaining a point cluster containing edge information in the image, and calculating the gradient consistency between the pixel point in the point cluster and other pixel points in the point cluster; step S13: setting a weight coefficient for gradient consistency, and determining the regularity of the pixel point with the weighted sum of multiple gradient consistencies as the overall trend; step S14: taking the ratio of the gradient regularity corresponding to the pixel point to the neighborhood grayscale chaos as the reliability of the pixel point; wherein, the calculation step of the gradient consistency of the pixel point includes: obtaining the gradient direction of each pixel point in the point cluster, and for each pixel point, obtaining the gradient direction difference and gradient value difference between the pixel point and other pixel points; obtaining gradient consistency based on the gradient direction difference and gradient value difference, and the gradient direction difference and gradient value difference are both inversely proportional to the gradient consistency.

2. The method for detecting the production quality of injection molded panels based on image enhancement according to claim 1, characterized in that: The steps for calculating the neighborhood grayscale disorder of a pixel point include: calculating the center difference of the pixel point based on the difference in grayscale values between the pixel point and the pixels in its corresponding neighborhood; calculating the neighborhood difference of the pixel point based on the difference in grayscale values of each pixel in the neighborhood, and taking the product of the neighborhood difference and the center difference as the neighborhood grayscale disorder of the pixel point.

3. The method for detecting the production quality of injection molded panels based on image enhancement according to claim 2, characterized in that: The formula for calculating the center difference is: Where, For the image OK The difference in the center of the pixels in the column; is the number of pixels in the neighborhood; Indicates the Rank Grayscale value of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The grayscale value extremes of the pixel in the column and all the pixels in its neighborhood.

4. The method for detecting the production quality of injection molded panels based on image enhancement according to claim 2, characterized in that: The formula for calculating neighborhood differences is: Where, Indicates the Rank Neighborhood differences of column pixels; Indicates the Rank The grayscale value extremes of the pixel in the column and all pixels in its neighborhood; For the Rank The number of pixels in the neighborhood of the column pixel; Indicates the Rank The neighborhood of the pixel in the column Gray value of each pixel; Indicates the Rank The neighborhood of the pixel in the column The grayscale value of a pixel.

5. The method for detecting the production quality of injection molded panels based on image enhancement according to claim 1, characterized in that: The step of obtaining a point cluster containing edge information in an image includes: obtaining a binary image of the image using an Otsu threshold method; and taking each white area in the binary image as a point cluster.

6. The method for detecting the production quality of injection molded panels based on image enhancement according to claim 1, characterized in that: The value normalized by the overall trend is the regularity of the pixel point.

7. The method for detecting production quality of injection molded panels based on image enhancement according to claim 6, characterized in that: The calculation formula of regularity is: Where, No. In the point cluster The regularity of pixels; For the The number of pixels in a cluster; Indicates the In the point cluster The overall trend of pixels.

8. The method for detecting the production quality of injection molded panels based on image enhancement according to claim 1, characterized in that: The calculation formula of the weight coefficient of the gradient consistency of the pixel points in the point cluster is: Where, Indicates the In the point cluster and The weight coefficient of the gradient consistency between pixels; Indicates the The first point in the cluster and The Euclidean distance between pixels.

9. The method for detecting production quality of injection molded panels based on image enhancement according to claim 1, characterized in that: The step of performing abnormality detection on the image based on the optimal image includes: setting an abnormal length; detecting an edge of the optimal image, and determining that a product has a defect in response to the edge length being greater than the abnormal length.

10. An injection molding panel production quality inspection system based on image enhancement, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an injection molding panel production quality detection method based on image enhancement according to any one of claims 1 to 9 is implemented.

Images