Quality Inspection Method, Equipment and Medium for Membrane-breaking Knife Based on Visual Detection

Abstract

The present invention discloses a method, device and medium for quality inspection of membrane-breaking knives based on visual inspection, belonging to the field of automated inspection technology. To solve the problems of low efficiency in quality inspection of membrane-breaking knives and difficulty in effectively counting the distribution positions of defects on membrane-breaking knives, the method based on comparison in a dot matrix comparison coordinate system ensures the accuracy and reliability of the comparison. By generating a standard edge contour line graph and a detected edge contour line graph and comparing them in the dot matrix comparison coordinate system, the wear condition and quality status of the membrane-breaking knife can be visually displayed, making full use of the high-precision characteristics of visual inspection technology. The defect distribution trend curve generated based on the defect distribution data group can visually display the distribution of defects on the membrane-breaking knife. This visual expression method not only facilitates the operator to quickly understand and identify the distribution law of defects, so as to make targeted improvements in the production process and quality control.

Description

Technical Field

[0001] The present invention relates to the field of automated detection technology, and in particular to a film-breaking knife quality detection method, equipment and medium based on visual detection. Background Art

[0002] Film breaking knife refers to the knife used in film blowing machine or other machinery for cutting or breaking film. With the continuous development of industrial manufacturing technology, film breaking knife is a key production tool, and its quality and performance directly affect production efficiency and product quality.

[0003] However, the detection methods under the existing technology still have the following problems in actual operation:

[0004] The traditional method of quality inspection of film-breaking knives often relies on manual visual inspection, which is not only inefficient but also easily affected by human factors, resulting in inaccurate and inconsistent test results. At the same time, it is difficult to effectively count the distribution of defects on the film-breaking knives during inspection, making it difficult to effectively optimize the production stage in a targeted manner. Summary of the invention

[0005] The purpose of the present invention is to provide a method, device and medium for detecting the quality of a membrane breaking knife based on visual detection, so as to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for detecting the quality of a membrane-breaking knife based on visual detection, the method specifically comprising:

[0007] Tool image acquisition, using the lighting device, imaging optical path and image acquisition unit to form a digital visual detection imaging component for the membrane-breaking knife, and using the digital visual detection imaging component to acquire the image of the blade of the membrane-breaking knife;

[0008] Tool image processing, preprocessing the tool edge image to obtain a preprocessed tool edge image;

[0009] Contour feature extraction: edge detection is performed on the pre-processed knife edge image and edge contour lines are drawn to obtain a detected edge contour line graph;

[0010] Tool quality inspection: create a dot matrix comparison coordinate system, bring the detected edge contour line graph and the standard edge contour line graph into the dot matrix comparison coordinate system, compare the detected edge contour line graph and the standard edge contour line graph through the dot matrix comparison coordinate system, and obtain a dot matrix comparison data set. The dot matrix comparison data set is compared through the preset allowable value threshold of the blade wear degree, and the quality of the film breaking knife is judged and classified. At the same time, a defect distribution trend curve is generated based on the dot matrix comparison data set.

[0011] Output the detection results, visually output the detection results of the film-breaking knife, generate a detection report, and at the same time, output the defect distribution trend curve.

[0012] Furthermore, the tool image acquisition specifically further includes the following steps:

[0013] The lighting device consists of a halogen lamp, a condenser lens, and a pinhole diaphragm. The imaging optical path consists of a light source system, a power supply, a beam splitter, a high-precision moving platform, and an imaging lens. The image acquisition unit consists of a CCD image sensor, a photoelectric converter, an image acquisition card, peripheral circuits, interfaces, and connecting wires;

[0014] Start the lighting device, place the film-breaking knife on the high-precision moving platform, and adjust the position and angle of the knife edge through the fine-tuning function of the platform. By adjusting the focal length and focusing position of the imaging lens, start the CCD image sensor, receive the optical image transmitted by the imaging optical path and convert it into an electrical signal. The photoelectric converter converts the electrical signal output by the CCD image sensor into a digital signal. The image acquisition card receives the digital signal output by the photoelectric converter, converts it, and outputs the knife-edge image.

[0015] Furthermore, the tool image processing specifically further includes the following steps:

[0016] Apply a filter to the acquired tool image to remove or reduce noise;

[0017] Apply the selected gray-scale transformation method to the filtered image to adjust the gray-scale distribution of the tool image and improve the contrast of the tool image;

[0018] Output the preprocessed knife-edge image.

[0019] Furthermore, adjusting the gray-scale distribution of the tool image and improving the contrast of the tool image includes:

[0020] Extract the tool image after removing or reducing noise;

[0021] Extract the gray-scale values corresponding to all pixel blocks in the tool image after removing or reducing noise;

[0022] Obtain a gray-scale reference value according to the gray-scale values corresponding to all pixel blocks in the tool image after removing or reducing noise; where the gray-scale reference value is obtained through the following formula;

[0023]

[0024] Where, H c represents the gray-scale reference value; H zrepresents the gray median corresponding to the gray values of all pixel blocks in the tool image after removing or reducing noise; n represents the number of pixel blocks; H i represents the gray value corresponding to the i-th pixel block;

[0025] Obtain the target gray value corresponding to each of the pixel blocks by combining the gray values corresponding to all the pixel blocks in the tool image after removing or reducing noise with a gray reference value;

[0026] Perform gray adjustment on the pixel blocks according to the target gray value corresponding to each of the pixel blocks.

[0027] Furthermore, obtaining the target gray value corresponding to each of the pixel blocks by combining the gray values corresponding to all the pixel blocks in the tool image after removing or reducing noise with a gray reference value includes:

[0028] Compare the gray values corresponding to all the pixel blocks in the tool image after removing or reducing noise with the gray reference value to obtain a gray value difference;

[0029] Extract the gray value differences corresponding to all the pixel blocks in the tool image after removing or reducing noise;

[0030] Obtain a difference factor by using the gray value differences corresponding to all the pixel blocks in the tool image after removing or reducing noise, where the difference factor is obtained through the following formula:

[0031]

[0032] where h represents the difference factor; n represents the number of pixel blocks; H i represents the gray value corresponding to the i-th pixel block; H z represents the gray median corresponding to the gray values of all pixel blocks in the tool image after removing or reducing noise; H fi represents the gray value difference corresponding to the i-th pixel block; h 01 and h 02 respectively represent a first adjustment coefficient and a second adjustment coefficient, and the first adjustment coefficient and the second adjustment coefficient are obtained through the following formula:

[0033]

[0034] where n represents the number of pixel blocks; H i represents the gray value corresponding to the i-th pixel block; H z represents the gray median corresponding to the gray values of all pixel blocks in the tool image after removing or reducing noise; H fi represents the gray value difference corresponding to the i-th pixel block;

[0035] The target gray level corresponding to the pixel block is obtained by combining the difference factor and the gray level difference corresponding to the pixel block with the gray level corresponding to the pixel block; wherein, the target gray level corresponding to the pixel block is obtained by the following formula:

[0036]

[0037] wherein, H m represents the target gray level corresponding to the pixel block; H 0 represents the gray level value corresponding to the pixel block; H f represents the gray level difference corresponding to the pixel block; H fp represents the average value corresponding to the gray level differences of all pixel blocks; h represents the difference factor.

[0038] Furthermore, the extraction of the contour features specifically further includes the following steps:

[0039] According to the preset feature information, edge detection is performed on the preprocessed knife-edge image, and several pixel points are extracted on the edge path of the preprocessed knife-edge image to obtain the edge points in the preprocessed knife-edge image;

[0040] Connect the detected edge points to form a continuous coil, and fit the coil to obtain the knife-edge contour line;

[0041] Draw the extracted knife-edge contour line on a blank image to obtain the original contour line diagram, and adjust the line width of the original contour line diagram, and output the detected edge contour line diagram after adjustment.

[0042] Furthermore, the detection of the tool quality specifically further includes obtaining a set of comparison data groups, including the following steps:

[0043] Obtain the standard drawing design drawing of the film-breaking knife, perform contour feature extraction on the film-breaking knife image in the standard drawing design drawing to obtain the standard edge contour line diagram;

[0044] Create a dot matrix comparison coordinate system, bring the standard edge contour line diagram into the dot matrix comparison coordinate system, make a horizontal reference line parallel to the x-axis based on the x-axis of the dot matrix comparison coordinate system, select the intersection point between the horizontal reference line and the standard edge contour line diagram as the standard intersection point, and obtain the coordinates (x 0 , y 0 ) of the standard intersection point in the dot matrix comparison coordinate system;

[0045] Move the horizontal reference line horizontally along the y-axis direction, obtain all the standard intersection points between the horizontal reference line and the standard edge contour line diagram during the movement, and use all the standard intersection points as the standard intersection point set;

[0046] Bring the detected edge contour line graph into the dot matrix comparison coordinate system. Make a horizontal reference line parallel to the x-axis based on the x-axis of the dot matrix comparison coordinate system. Select the intersection point between the horizontal reference line and the detected edge contour line graph as the detection intersection point, and obtain the coordinates (a 0 , b 0 ) of the detection intersection point in the dot matrix comparison coordinate system;

[0047] Horizontally move the horizontal reference line along the y-axis direction to obtain all detection intersection points between the horizontal reference line and the detected edge contour line graph during the movement, and use all detection intersection points as the detection intersection point set;

[0048] Perform correlation comparison between the standard intersection point set and the detection intersection point set;

[0049] Select each standard intersection point in the standard intersection point set and obtain the x in the coordinates of each standard intersection point 0 , select each detection intersection point in the detection intersection point set and obtain the a in the coordinates of each detection intersection point 0 . Correlate the standard intersection points and detection intersection points with the same x 0 and a 0 values to generate associated intersection point groups, and obtain the y 0 and b 0 in the associated intersection point groups;

[0050] Calculate the difference T 0 between y 0 and b 0 in the associated intersection point groups. Combine the a 0 in the coordinates of each detection intersection point with the difference T 0 corresponding to the associated intersection point group to which the detection intersection point belongs to obtain a comparison data group (a 0 , T 0 ). Obtain the comparison data group of each detection intersection point in the detection intersection point set to obtain a dot matrix comparison data group set.

[0051] Furthermore, the tool quality detection specifically further includes cutting edge wear comparison and trend curve generation, including the following steps:

[0052] Generate an allowable value threshold for the preset cutting edge wear degree based on the actual production standard, and the allowable value threshold includes a first preset threshold and a second preset threshold;

[0053] Obtain the difference T 0 in all comparison data groups in the comparison data group set and generate a threshold detection data set. Perform threshold comparison on each difference T 0 in the threshold detection data set through the first preset threshold and the second preset threshold;

[0054] If there is a difference T greater than or equal to the first preset threshold in the threshold detection dataset 0 , it is determined that the quality of the membrane-breaking knife corresponding to the threshold detection dataset is unqualified;

[0055] If there is a difference T in the threshold detection dataset that is less than the first preset threshold and greater than the second preset threshold 0 , it is determined that the quality of the membrane-breaking knife corresponding to the threshold detection dataset needs to be repaired;

[0056] Otherwise, it is determined that the quality of the membrane-breaking knife corresponding to the threshold detection dataset is qualified;

[0057] Obtain the set of comparison data groups corresponding to all membrane-breaking knives with unqualified and to-be-repaired qualities as the defect set, and select the comparison data groups in the defect set where the difference T 0 is not 0 as the defect data groups, and obtain the a in the defect data groups 0 , and integrate the a in all defect data groups 0 and generate a defect distribution data group;

[0058] Based on the a in the defect distribution data group 0 generate a defect distribution trend curve.

[0059] A membrane-breaking knife quality detection device based on visual detection includes a processor, an input device, an output device, and a memory. The processor, input device, output device, and memory are connected in sequence. The memory is used to store a computer program. The computer program includes program instructions. The processor is configured to call the program instructions to execute the above-mentioned membrane-breaking knife quality detection method based on visual detection.

[0060] Furthermore, a computer-readable storage medium stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, the processor executes the above-mentioned membrane-breaking knife quality detection method based on visual detection.

[0061] Compared with the prior art, the beneficial effects of the present invention are:

[0062] 1. The present invention can accurately locate the wear degree and quality of the film-breaking knife by extracting the contour features of the film-breaking knife image in the standard drawing design and comparing them with the actually detected film-breaking knife image. This method based on the dot matrix comparison coordinate system ensures the accuracy and reliability of the comparison. By setting the preset allowable value threshold for the wear degree of the knife edge, the quality of the film-breaking knife can be quickly and accurately determined, which not only improves the efficiency of quality inspection but also effectively identifies unqualified and repairable film-breaking knives. By generating the standard edge contour line graph and the detected edge contour line graph and comparing them in the dot matrix comparison coordinate system, the wear condition and quality status of the film-breaking knife can be visually displayed, which helps the operator quickly understand and identify the problem and provides convenience for subsequent processing and repair.

[0063] 2. The present invention can accurately locate the defect positions on the film-breaking knife by collecting the comparison data groups of unqualified and repairable film-breaking knives and screening the defect data groups, making full use of the high-precision characteristics of the visual detection technology and effectively avoiding the omissions and errors that may exist in traditional manual detection. The defect distribution trend curve generated based on the defect distribution data group can visually display the distribution of defects on the film-breaking knife. This visual expression method not only facilitates the operator to quickly understand and identify the distribution law of defects but also provides strong support for subsequent quality improvement and process optimization. By analyzing the defect distribution trend curve, it can be judged which parts of the film-breaking knife are prone to defects, so as to improve the production process and quality control targeted. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 is the flow chart of the detection method of the present invention;

[0065] Figure 2 is the schematic diagram of the film-breaking knife quality detection method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0066] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0067] To solve the technical problems that the traditional quality detection method of the film-breaking knife often relies on manual visual inspection, which is not only inefficient but also easily affected by human factors, resulting in inaccurate and inconsistent detection results, and it is difficult to effectively count the positions of defect distributions on the film-breaking knife during detection and it is difficult to optimize the production stage targeted, please refer toFigure 1-2 , the present invention provides the following technical solutions:

[0068] A quality inspection method for a film-breaking knife based on visual inspection, which specifically includes:

[0069] Tool image acquisition: Using an illumination device, an imaging optical path, and an image acquisition unit to form a digital visual inspection imaging component for the film-breaking knife, and collecting the edge image of the film-breaking knife through the digital visual inspection imaging component;

[0070] Tool image processing: Preprocessing the edge image to obtain a preprocessed edge image;

[0071] Contour feature extraction: Performing edge detection on the preprocessed edge image and drawing an edge contour line to obtain a detected edge contour line diagram;

[0072] Tool quality inspection: Creating a dot matrix comparison coordinate system, bringing the detected edge contour line diagram and the standard edge contour line diagram into the dot matrix comparison coordinate system, comparing the detected edge contour line diagram and the standard edge contour line diagram through the dot matrix comparison coordinate system to obtain a set of dot matrix comparison data groups, comparing the set of dot matrix comparison data groups through a preset allowable value threshold for the degree of tool edge wear, judging and classifying the quality of the film-breaking knife, and simultaneously generating a defect distribution trend curve based on the set of dot matrix comparison data groups;

[0073] Detection result output: Outputting the detection result of the film-breaking knife in a visual manner to generate a detection report, and at the same time, outputting the defect distribution trend curve.

[0074] In the above embodiment, through the digital visual inspection imaging component, the edge image of the film-breaking knife can be accurately collected. The collaborative work of the illumination device, the imaging optical path, and the image acquisition unit ensures the clarity and accuracy of the image acquisition, providing a reliable data basis for subsequent image processing and quality inspection. The entire inspection method realizes automation and intelligence, reduces the interference of human factors, improves the inspection efficiency and accuracy. Through the preset algorithms and thresholds, the system can automatically judge and classify the quality of the film-breaking knife, reducing the tediousness of manual operations.

[0075] In the above embodiment, the detection result can be output in a visual manner, enabling the operator to intuitively understand the quality status of the film-breaking knife. At the same time, the generated detection report and defect distribution trend curve are helpful for tracking and analyzing the tool wear situation, providing data support for the optimization of the production process. By adjusting the dot matrix comparison coordinate system and the preset allowable value threshold, this inspection method can adapt to the quality inspection requirements of different specifications and types of film-breaking knives.

[0076] Tool image acquisition specifically further includes the following steps:

[0077] The lighting device consists of a halogen lamp, a condenser lens, and a pinhole diaphragm. The imaging optical path consists of a light source system, a power supply, a beam splitter, a high-precision moving platform, and an imaging lens. The image acquisition unit consists of a CCD image sensor, a photoelectric converter, an image acquisition card, peripheral circuits, interfaces, and connecting wires;

[0078] Start the lighting device, place the film-breaking knife on the high-precision moving platform, and adjust the position and angle of the knife edge through the fine-tuning function of the platform. By adjusting the focal length and focusing position of the imaging lens, start the CCD image sensor to receive the optical image transmitted by the imaging optical path and convert it into an electrical signal. The photoelectric converter converts the electrical signal output by the CCD image sensor into a digital signal. The image acquisition card receives the digital signal output by the photoelectric converter, converts it, and outputs the image of the knife edge.

[0079] In the above embodiment, by starting the lighting device and using the combination of the halogen lamp, the condenser lens, and the pinhole diaphragm, uniform and high-brightness illumination of the knife edge of the film-breaking knife can be ensured. The fine-tuning function of the high-precision moving platform enables the operator to precisely adjust the position and angle of the film-breaking knife, ensuring that the imaging lens can capture the best image. By adjusting the focal length and focusing position of the imaging lens, it can be ensured that the image acquisition unit receives clear and accurate images.

[0080] In the above embodiment, the CCD image sensor can receive the optical image transmitted by the imaging optical path and convert it into an electrical signal. The photoelectric converter further converts these electrical signals into digital signals to ensure the digital representation of the image data. The acquisition of the tool image is one of the key steps in the quality inspection method of the film-breaking knife based on visual inspection and is closely coordinated with other steps (such as image processing, contour feature extraction, quality inspection, etc.). This coordination ensures the coherence and efficiency of the entire inspection process and improves the accuracy and reliability of the inspection results.

[0081] The image processing of the tool specifically further includes the following steps:

[0082] Apply a filter to the acquired tool image to remove or reduce noise;

[0083] Apply the selected gray-scale transformation method to the filtered image to adjust the gray-scale distribution of the tool image and improve the contrast of the tool image;

[0084] Output the preprocessed knife-edge image.

[0085] In the above embodiments, by applying a filter, the noise in the tool image can be effectively removed or reduced, and the interference factors in the image can be decreased. By applying the selected gray-scale transformation method, the gray-scale distribution of the tool image can be adjusted, and the contrast of the image can be enhanced. The preprocessed tool-edge image after filtering and gray-scale transformation not only removes the noise and enhances the contrast but also retains the key feature information of the tool, which provides a good image basis for subsequent steps such as contour feature extraction, edge detection, and quality inspection, ensuring the accuracy and reliability of the entire inspection process.

[0086] In the above embodiments, through automated image processing techniques, the preprocessing of the tool image can be completed quickly and accurately, avoiding the cumbersome process and errors of traditional manual processing methods, helping to improve the inspection efficiency, reducing the inspection errors caused by human factors at the same time, and enhancing the accuracy and reliability of the entire quality inspection process.

[0087] Specifically, adjusting the gray-scale distribution of the tool image and enhancing the contrast of the tool image includes:

[0088] Extracting the tool image after removing or reducing the noise;

[0089] Extracting the gray-scale values corresponding to all pixel blocks in the tool image after removing or reducing the noise;

[0090] Obtaining a gray-scale reference value according to the gray-scale values corresponding to all pixel blocks in the tool image after removing or reducing the noise; wherein, the gray-scale reference value is obtained through the following formula;

[0091]

[0092] wherein, H c represents the gray-scale reference value; H z represents the median gray-scale value corresponding to the gray-scale values of all pixel blocks in the tool image after removing or reducing the noise; n represents the number of pixel blocks; H i represents the gray-scale value corresponding to the i-th pixel block;

[0093] Obtaining the target gray-scale corresponding to each pixel block by combining the gray-scale values corresponding to all pixel blocks in the tool image after removing or reducing the noise with the gray-scale reference value;

[0094] Adjusting the gray-scale of each pixel block according to the target gray-scale corresponding to each pixel block.

[0095] The technical effects of the above technical solution are as follows: By removing or reducing noise, the initial quality of the tool image is first improved. Noise is the unnecessary random variation in the image, which is used to obscure the important details or features of the image and reduce the clarity of the image. After removing or reducing noise, the image becomes cleaner and clearer, providing a better basis for subsequent processing.

[0096] By calculating the gray reference value (Hc) and adjusting the gray value of each pixel block based on this value, the technical solution realizes the enhancement of the contrast of the tool image. Contrast is the degree of difference between the brightest part and the darkest part in the image. An image with high contrast can more clearly display the outline and details of an object. This adjustment makes features such as edges and textures in the tool image more prominent, facilitating subsequent image analysis and processing.

[0097] The calculation of the gray reference value takes into account the gray values of all pixel blocks in the image, especially by the weighted average of the gray median (Hz) and the gray value of the pixel block (Hi). This method makes the gray adjustment more adaptable to the actual content of the image, avoiding the overexposure or underexposure problems that may be caused by a single threshold adjustment, and ensuring the natural transition and detail retention of the image.

[0098] The tool image after gray distribution adjustment and contrast enhancement will have higher accuracy in subsequent image analysis (such as edge detection, shape recognition, defect detection, etc.). This is because the increased contrast and reduced noise in the image make the features in the image more distinct, reducing the possibility of misjudgment and missed judgment. At the same time, due to the improved quality of the processed image, the amount of calculation and time required in subsequent processing steps may be reduced, thus improving the overall processing efficiency.

[0099] In summary, the technical solution significantly improves the quality of the image and the accuracy of subsequent analysis by adjusting the gray distribution and contrast of the tool image, and has significant technical effects and application value.

[0100] Specifically, obtaining the target gray value corresponding to each of the pixel blocks by combining the gray values corresponding to all the pixel blocks in the tool image after removing or reducing noise with the gray reference value includes:

[0101] Comparing the gray values corresponding to all the pixel blocks in the tool image after removing or reducing noise with the gray reference value to obtain the gray value difference;

[0102] Extracting the gray value differences corresponding to all the pixel blocks in the tool image after removing or reducing noise;

[0103] Obtaining a difference factor by using the gray value differences corresponding to all the pixel blocks in the tool image after removing or reducing noise, where the difference factor is obtained by the following formula:

[0104]

[0105] Among them, h represents the difference factor; n represents the number of pixel blocks; H i represents the gray value corresponding to the i-th pixel block; H z represents the gray median corresponding to the gray values of all pixel blocks in the tool image after removing or reducing noise; H fi represents the difference in gray value corresponding to the i-th pixel block; h 01 and h 02 respectively represent the first adjustment coefficient and the second adjustment coefficient, and the first adjustment coefficient and the second adjustment coefficient are obtained through the following formula:

[0106]

[0107] Among them, n represents the number of pixel blocks; H i represents the gray value corresponding to the i-th pixel block; H z represents the gray median corresponding to the gray values of all pixel blocks in the tool image after removing or reducing noise; H fi represents the difference in gray value corresponding to the i-th pixel block;

[0108] Using the difference factor and the difference in gray value corresponding to the pixel block, combined with the gray value corresponding to the pixel block, to obtain the target gray value corresponding to the pixel block; among them, the target gray value corresponding to the pixel block is obtained through the following formula:

[0109]

[0110] Among them, H m represents the target gray value corresponding to the pixel block; H 0 represents the gray value corresponding to the pixel block; H f represents the difference in gray value corresponding to the pixel block; H fp represents the average value corresponding to the difference in gray values of all pixel blocks; h represents the difference factor.

[0111] The technical effect of the above technical solution is that by comparing the gray value of each pixel block with the gray reference value and calculating the difference in gray value, this technical solution can more finely analyze the deviation degree of each pixel block in the image from the overall gray distribution. This refined analysis provides a more accurate basis for the subsequent calculation of the target gray value.

[0112] The difference factor (h) is introduced as the basis for adjusting the grayscale value. This factor is calculated by comprehensively considering the grayscale value, the median grayscale value, the grayscale value difference of the pixel block, and two adjustment coefficients (h01 and h02). This design enables the grayscale adjustment process to adapt to the specific content of the image, avoiding a one-size-fits-all approach, thus preserving important details and features in the image.

[0113] When calculating the difference factor, the grayscale value differences of all pixel blocks are considered, and two adjustment coefficients (h01 and h02) are used to balance the influence of the grayscale value difference on the difference factor. This balanced design helps prevent adjustment deviations caused by individual extreme grayscale values, ensuring the stability and consistency of the overall adjustment effect.

[0114] By using the difference factor and the grayscale value difference corresponding to the pixel block to calculate the target grayscale and adjusting the grayscale of the pixel block accordingly, this technical solution can more effectively enhance the contrast of the image, making the details of the image clearer while maintaining the natural transition and overall visual effect of the image.

[0115] The tool image after the above grayscale adjustment will be more accurate in the subsequent image analysis process. Because features such as edges and textures in the image are more prominent, reducing the possibility of misjudgment and missed judgment, thus improving the efficiency and reliability of image analysis.

[0116] The adjustment coefficients (h01 and h02) in this technical solution can be adjusted according to specific application scenarios and requirements to optimize the grayscale adjustment effect. This flexibility enables this technical solution to adapt to different types of tool images and different processing requirements.

[0117] In summary, this technical solution significantly improves the quality of the tool image and the accuracy of subsequent analysis through refined grayscale value analysis and adaptive grayscale adjustment strategies, having significant technical effects and broad application prospects.

[0118] Contour feature extraction specifically further includes the following steps:

[0119] According to the preset feature information, edge detection is performed on the preprocessed knife-edge image, and several pixel points are extracted on the edge path of the preprocessed knife-edge image to obtain the edge points in the preprocessed knife-edge image;

[0120] Connect the detected edge points to form a continuous coil, and fit the coil to obtain the knife-edge contour line;

[0121] Draw the extracted knife-edge contour line on a blank image to obtain the original contour line diagram, and adjust the line width of the original contour line diagram, and output the detected edge contour line diagram after adjustment.

[0122] In the above embodiments, through the preset feature information, this technical solution can accurately perform edge detection on the preprocessed tool edge image, effectively extract the key pixel points of the tool edge, connect the detected edge points to form a continuous coil, and perform fitting processing on it, so as to obtain a smooth and accurate tool edge contour line. This not only removes possible noise points, but also ensures the continuity and integrity of the contour line, providing a reliable basis for subsequent quality analysis and evaluation.

[0123] In the above embodiments, the extracted tool edge contour line is drawn on a blank image to form an original contour line diagram, which can visually display the shape and features of the tool edge. By adjusting the line width of the original contour line diagram, the display effect of the contour line can be further optimized. Appropriate line width adjustment can make the contour line clearer and more prominent, helping to improve the accuracy and efficiency of visual inspection.

[0124] In the above embodiments, through the automated contour feature extraction technology, the contour information of the tool edge can be quickly and accurately obtained, avoiding the cumbersome process and errors of traditional manual measurement methods, helping to improve the efficiency of quality inspection, while reducing the detection errors caused by human factors and enhancing the accuracy and reliability of the entire quality inspection process.

[0125] The tool quality inspection specifically further includes the following steps:

[0126] Obtain the standard drawing design of the film-breaking tool, perform contour feature extraction based on the film-breaking tool image in the standard drawing design to obtain a standard edge contour line diagram;

[0127] Create a dot matrix comparison coordinate system, bring the standard edge contour line diagram into the dot matrix comparison coordinate system, make a horizontal reference line parallel to the x-axis based on the x-axis of the dot matrix comparison coordinate system, select the intersection point between the horizontal reference line and the standard edge contour line diagram as the standard intersection point, and obtain the coordinates (x 0 , y 0 ) of the standard intersection point in the dot matrix comparison coordinate system;

[0128] Move the horizontal reference line horizontally along the y-axis direction to obtain all the standard intersection points between the horizontal reference line and the standard edge contour line diagram during the movement, and use all the standard intersection points as the standard intersection point set;

[0129] Bring the detected edge contour line diagram into the dot matrix comparison coordinate system, make a horizontal reference line parallel to the x-axis based on the x-axis of the dot matrix comparison coordinate system, select the intersection point between the horizontal reference line and the detected edge contour line diagram as the detected intersection point, and obtain the coordinates (a 0 , b 0 ) of the detected intersection point in the dot matrix comparison coordinate system;

[0130] Horizontally move the horizontal reference line along the y-axis direction to obtain all the detection intersection points between the horizontal reference line and the detected edge contour line graph during the movement, and use all the detection intersection points as the detection intersection point set;

[0131] Perform an association comparison between the standard intersection point set and the detection intersection point set;

[0132] Select each standard intersection point in the standard intersection point set and obtain the x in the coordinates of each standard intersection point 0 , select each detection intersection point in the detection intersection point set and obtain the a in the coordinates of each detection intersection point 0 , for x 0 and a 0 Associate the standard intersection point and the detection intersection point with the same numerical value of x and a to generate an associated intersection point group, and obtain the y in the associated intersection point group 0 and b 0 ;

[0133] Calculate the difference T between y 0 and b 0 in the associated intersection point group, and combine the a 0 in the coordinates of each detection intersection point with the difference T 0 corresponding to the associated intersection point group to which the detection intersection point belongs to obtain a comparison data group (a 0 , T 0 , T 0 ), obtain the comparison data group of each detection intersection point in the detection intersection point set to obtain a dot matrix comparison data group set.

[0134] Generate an allowable value threshold for the preset knife edge wear degree based on the actual production standard, and the allowable value threshold includes a first preset threshold and a second preset threshold;

[0135] Obtain the difference T 0 in all the comparison data groups in the comparison data group set and generate a threshold detection data set, and perform a threshold comparison on each difference T 0 in the threshold detection data set through the first preset threshold and the second preset threshold;

[0136] If there is a difference T greater than or equal to the first preset threshold in the threshold detection data set 0 , determine that the quality of the film-breaking knife corresponding to the threshold detection data set is unqualified;

[0137] If there is a difference T less than the first preset threshold and greater than the second preset threshold in the threshold detection data set 0 , determine that the quality of the film-breaking knife corresponding to the threshold detection data set needs to be repaired;

[0138] Otherwise, it is determined that the quality of the membrane-breaking knife corresponding to the threshold detection data set is qualified.

[0139] In the above embodiment, by extracting the contour features of the membrane-breaking knife image in the standard drawing design drawing and comparing them with the actually detected membrane-breaking knife image, the wear degree and quality of the membrane-breaking knife can be accurately located. This method based on the dot matrix comparison coordinate system ensures the accuracy and reliability of the comparison and provides a solid data basis for the subsequent threshold comparison.

[0140] In the above embodiment, by setting the preset allowable value thresholds for the edge wear degree of the knife (the first preset threshold and the second preset threshold), the quality of the membrane-breaking knife can be quickly and accurately determined, which not only improves the efficiency of quality inspection but also effectively identifies unqualified and knives to be repaired, providing strong support for quality control in the production process.

[0141] In the above embodiment, by generating the standard edge contour line graph and the detected edge contour line graph and comparing them in the dot matrix comparison coordinate system, the wear condition and quality status of the membrane-breaking knife can be intuitively displayed, which helps the operator quickly understand and identify the problems and provides convenience for subsequent processing and repair.

[0142] Obtain the set of comparison data groups corresponding to all membrane-breaking knives with unqualified quality and those to be repaired as the defect set, and select the comparison data groups with non-zero difference T 0 in the defect set as the defect data groups, and obtain the a 0 in the defect data groups, and integrate the a 0 in all defect data groups and generate the defect distribution data group;

[0143] Based on the a 0 in the defect distribution data group, generate the defect distribution trend curve.

[0144] In the above embodiment, by collecting the comparison data groups of membrane-breaking knives with unqualified quality and those to be repaired and screening out the defect data groups with non-zero difference T 0 in them, the defect positions existing on the membrane-breaking knife can be accurately located, making full use of the high-precision characteristics of the visual detection technology and effectively avoiding the omissions and errors that may exist in traditional manual detection. By integrating the a 0 values in all defect data groups, a defect distribution data group can be formed, and this data group provides the necessary data basis for generating the defect distribution trend curve subsequently.

[0145] In the above embodiments, the defect distribution trend curve generated based on the defect distribution data group can intuitively display the distribution of defects on the film-breaking knife. This visual expression method not only facilitates the operator to quickly understand and identify the distribution law of defects, but also provides strong support for subsequent quality improvement and process optimization. By analyzing the defect distribution trend curve, it can be judged which parts of the film-breaking knife are prone to defects, so as to improve the production process and quality control targeted, which helps to reduce the number of unqualified products and products to be repaired, improve production efficiency, and also helps to improve the overall quality level of the film-breaking knife.

[0146] The film-breaking knife quality detection device based on visual detection includes a processor, an input device, an output device and a memory. The processor, the input device, the output device and the memory are connected in sequence. The memory is used to store a computer program, and the computer program includes program instructions. The processor is configured to call the program instructions to execute the above-mentioned film-breaking knife quality detection method based on visual detection.

[0147] A computer-readable storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by a processor, the processor is caused to execute the above-mentioned film-breaking knife quality detection method based on visual detection.

[0148] As mentioned above, the above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent replacements or changes, and all should be covered by the protection scope of the present invention.

Claims

1. The film breaking knife quality detection method based on visual detection is characterized by: Specifically include: Tool image acquisition, using the lighting device, imaging optical path and image acquisition unit to form a digital visual detection imaging component for the membrane-breaking knife, and using the digital visual detection imaging component to acquire the image of the blade of the membrane-breaking knife; Tool image processing, preprocessing the tool edge image to obtain a preprocessed tool edge image; Contour feature extraction: edge detection is performed on the pre-processed knife edge image and edge contour lines are drawn to obtain a detected edge contour line graph; Tool quality inspection: create a dot matrix comparison coordinate system, bring the detected edge contour line graph and the standard edge contour line graph into the dot matrix comparison coordinate system, compare the detected edge contour line graph and the standard edge contour line graph through the dot matrix comparison coordinate system, and obtain a dot matrix comparison data set. The dot matrix comparison data set is compared through the preset allowable value threshold of the blade wear degree, and the quality of the film breaking knife is judged and classified. At the same time, a defect distribution trend curve is generated based on the dot matrix comparison data set. Output of test results: Output the test results of the film-breaking knife in a visual way, generate a test report, and output the defect distribution trend curve at the same time; The tool image processing further includes the following steps: Applying a filter to the acquired tool image to remove or attenuate noise; Applying the selected grayscale transformation method to the filtered image, adjusting the grayscale distribution of the tool image, and improving the contrast of the tool image; Output preprocessed knife edge image; Among them, adjusting the grayscale distribution of the tool image and improving the contrast of the tool image include: Extracting a tool image after removing or reducing noise; Extracting the grayscale values ​​corresponding to all pixel blocks in the tool image after the noise is removed or weakened; Obtain a grayscale reference value according to the grayscale values ​​corresponding to all pixel blocks in the tool image after the noise is removed or weakened; wherein the grayscale reference value is obtained by the following formula; Among them, H c Indicates the grayscale reference value; H z represents the grayscale median value corresponding to the grayscale values ​​of all pixel blocks in the tool image after the noise is removed or weakened; n represents the number of pixel blocks; H i Represents the grayscale value corresponding to the i-th pixel block; Obtain a target grayscale corresponding to each pixel block by using the grayscale values ​​corresponding to all pixel blocks in the tool image after the noise is removed or weakened in combination with a grayscale reference value; The grayscale of each pixel block is adjusted according to the target grayscale corresponding to each pixel block.

2. The method for detecting the quality of a film-breaking knife based on visual detection according to claim 1, characterized in that: The tool image acquisition specifically includes the following steps: The lighting device is composed of a halogen lamp, a condenser and a pinhole diaphragm, the imaging optical path is composed of a light source system, a power supply, a spectroscope, a high-precision mobile platform and an imaging lens, and the image acquisition unit is composed of a CCD image sensor, a photoelectric converter, an image acquisition card, a peripheral circuit, an interface and a connecting line; Start the lighting device, place the membrane breaking knife on the high-precision mobile platform, and adjust the position and angle of the blade through the fine-tuning function of the platform. By adjusting the focal length and focus position of the imaging lens, start the CCD image sensor, receive the optical image transmitted by the imaging light path and convert it into an electrical signal, the photoelectric converter converts the electrical signal output by the CCD image sensor into a digital signal, and the image acquisition card receives the digital signal output by the photoelectric converter, converts and outputs the blade image.

3. The method for detecting the quality of a membrane-breaking knife based on visual detection according to claim 1, characterized in that: The target grayscale corresponding to each pixel block is obtained by using the grayscale values ​​corresponding to all pixel blocks in the tool image after the noise is removed or weakened in combination with the grayscale reference value, including: Compare the grayscale values ​​corresponding to all pixel blocks in the tool image after the noise is removed or weakened with the grayscale reference value to obtain a grayscale value difference; Extracting grayscale value differences corresponding to all pixel blocks in the tool image after the noise is removed or reduced; The difference factor is obtained by using the gray value difference corresponding to all pixel blocks in the tool image after the noise is removed or weakened, wherein the difference factor is obtained by the following formula: Where h is the difference factor; n is the number of pixel blocks; H i Represents the grayscale value corresponding to the i-th pixel block; H z represents the grayscale median value corresponding to the grayscale values ​​of all pixel blocks in the tool image after the noise is removed or weakened; H fi represents the gray value difference corresponding to the i-th pixel block; h 01 and h 02 represent the first adjustment coefficient and the second adjustment coefficient respectively, and the first adjustment coefficient and the second adjustment coefficient are obtained by the following formula: Where n is the number of pixel blocks; H i Represents the grayscale value corresponding to the i-th pixel block; H z represents the grayscale median value corresponding to the grayscale values ​​of all pixel blocks in the tool image after the noise is removed or weakened; H fi Represents the gray value difference corresponding to the i-th pixel block; The target grayscale corresponding to the pixel block is obtained by using the difference factor and the grayscale value difference corresponding to the pixel block in combination with the grayscale value corresponding to the pixel block; wherein the target grayscale corresponding to the pixel block is obtained by the following formula: Among them, H m represents the target grayscale corresponding to the pixel block; H0 represents the grayscale value corresponding to the pixel block; H f Indicates the gray value difference corresponding to the pixel block; H fp It represents the average value corresponding to the gray value difference of all pixel blocks; h represents the difference factor.

4. The method for detecting the quality of a membrane-breaking knife based on visual detection according to claim 1, characterized in that: The contour feature extraction specifically includes the following steps: According to the preset feature information, edge detection is performed on the pre-processed knife edge image, and a number of pixel points are extracted on the edge path of the pre-processed knife edge image to obtain edge points in the pre-processed knife edge image; Connecting the detected edge points to form a continuous coil, and fitting the coil to obtain a knife edge contour line; The extracted edge contour line is drawn on a blank image to obtain an original contour line map, and the line width of the original contour line map is adjusted. After the adjustment, a detection edge contour line map is output.

5. The method for detecting the quality of a membrane-breaking knife based on visual detection according to claim 1, characterized in that: The tool quality detection specifically includes obtaining a comparison data set, including the following steps: Obtain a standard design drawing of a membrane-breaking knife, extract contour features based on the membrane-breaking knife image in the standard design drawing, and obtain a standard edge contour line drawing; Create a dot matrix comparison coordinate system, bring the standard edge contour line graph into the dot matrix comparison coordinate system, make a horizontal reference line parallel to the x-axis based on the x-axis of the dot matrix comparison coordinate system, select the intersection point between the horizontal reference line and the standard edge contour line graph as the standard intersection point, and obtain the coordinates (x0, y0) of the standard intersection point in the dot matrix comparison coordinate system; The horizontal reference straight line is moved horizontally along the y-axis direction, and all standard intersection points between the horizontal reference straight line and the standard edge contour line graph during the movement are obtained, and all the standard intersection points are taken as a standard intersection point set; Bring the detection edge contour line graph into the dot matrix comparison coordinate system, make a horizontal reference line parallel to the x-axis based on the x-axis of the dot matrix comparison coordinate system, select the intersection point between the horizontal reference line and the detection edge contour line graph as the detection intersection point, and obtain the coordinates (a0, b0) of the detection intersection point in the dot matrix comparison coordinate system; The horizontal reference straight line is moved horizontally along the y-axis direction, and all the detection intersection points between the horizontal reference straight line and the detection edge contour line graph during the movement are obtained, and all the detection intersection points are taken as a detection intersection point set; Perform correlation comparison between the standard intersection point set and the detection intersection point set; Select each standard intersection point in the standard intersection point set, obtain x0 in the coordinates of each standard intersection point, select each detection intersection point in the detection intersection point set, obtain a0 in the coordinates of each detection intersection point, associate the standard intersection points and detection intersection points with the same x0 and a0 values, generate an associated intersection point group, and obtain y0 and b0 in the associated intersection point group; Calculate the difference T0 between y0 and b0 in the associated intersection group, combine a0 in the coordinates of each detected intersection with the difference T0 corresponding to the associated intersection group to which the detected intersection belongs, and obtain a comparison data group (a0, T0), obtain the comparison data group of each detected intersection in the detected intersection set, and obtain a dot matrix comparison data group set.

6. The method for detecting the quality of a membrane-breaking knife based on visual detection according to claim 5, characterized in that: The tool quality detection specifically includes tool edge wear comparison and trend curve generation, including the following steps: Generate a preset allowable value threshold of the blade wear degree based on the actual production standard, wherein the allowable value threshold includes a first preset threshold and a second preset threshold; Obtaining the difference values ​​T0 in all the comparison data groups in the comparison data group set and generating a threshold detection data set, and performing a threshold comparison on each difference value T0 in the threshold detection data set by using a first preset threshold value and a second preset threshold value; If there is a difference T0 greater than or equal to the first preset threshold value in the threshold detection data set, it is determined that the quality of the membrane breaking knife corresponding to the threshold detection data set is unqualified; If there is a difference T0 in the threshold detection data set that is smaller than the first preset threshold and larger than the second preset threshold, it is determined that the quality of the membrane-breaking knife corresponding to the threshold detection data set is to be repaired; Otherwise, it is determined that the membrane breaking knife corresponding to the threshold detection data set is of qualified quality; Obtain a set of comparison data groups corresponding to all film-breaking knives with unqualified quality and to-be-repaired as a defect set, select a comparison data group whose difference T0 is not 0 in the defect set as a defect data group, obtain a0 in the defect data group, integrate a0 in all defect data groups and generate a defect distribution data group; A defect distribution trend curve is generated based on a0 in the defect distribution data set.

7. The film breaking knife quality inspection equipment based on visual inspection is characterized by: It includes a processor, an input device, an output device and a memory, which are connected in sequence. The memory is used to store a computer program. The computer program includes program instructions. The processor is configured to call the program instructions to execute the membrane breaking knife quality detection method based on visual inspection as described in claims 1-6.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, which includes program instructions. When the program instructions are executed by the processor, the processor executes the membrane breaking knife quality detection method based on visual detection as described in claims 1-6.

Images