Glass production quality detection system based on machine vision

Abstract

The invention relates to the technical field of glass quality detection, in particular to a glass production quality detection system based on machine vision, which comprises an image acquisition module, an illumination optimization analysis module, an image processing module, a control module, a feedback report module and a data storage end. According to the method, the illumination state information is obtained by analyzing the illumination data of each sub-region of the illumination region, the illumination abnormal region is obtained according to the illumination information, the corresponding illumination adjustment information is obtained according to the illumination abnormal region, and the illumination information of the abnormal region can be adjusted through the illumination adjustment information; the glass quality detection effect is prevented from being reduced by abnormal illumination; the glass quality information is obtained by analyzing the product image data, the glass production abnormity report is obtained according to the glass quality information, and the control failure report is generated according to the control effect, so that the abnormity information can be checked conveniently.

Description

Technical Field

[0001] The present invention relates to the technical field of glass quality detection, and in particular to a glass production quality detection system based on machine vision. Background Art

[0002] As an important building material and industrial material, glass is widely used in the fields of construction, automobiles, electronics, optics, etc. Different fields have different requirements for glass quality. For example, automotive glass needs to have high strength and transparency, and electronic glass needs to have good flatness and conductivity. Therefore, glass manufacturers need a more stringent quality inspection system to ensure product quality. The existing machine vision-based glass production quality inspection system needs to collect glass images through a camera when in use. However, when collecting images, uneven lighting can easily cause unstable image quality, affect feature extraction and defect detection, and reduce the quality inspection effect of glass production. Summary of the invention

[0003] The present invention provides a glass production quality detection system based on machine vision, which is used to solve the above technical problems.

[0004] A first aspect of the present invention provides a glass production quality inspection system based on machine vision, including an image acquisition module, a lighting optimization analysis module, an image processing module, a control module, a feedback report module and a data storage terminal.

[0005] The image acquisition module collects light data through the light sensor, and uses a high-resolution industrial camera to capture the product image of the glass production line in real time to obtain product image data, and sends the product image data and light data to the data storage end; The lighting data includes lighting environment data, light source array data and optical characteristic data; the product image data includes glass size data, surface quality information and cross-sectional image data.

[0006] The lighting optimization and analysis module is used to divide the lighting area into multiple sub-areas, obtain the lighting data of each sub-area and analyze the lighting data to obtain lighting status information, identify the lighting status information to obtain abnormal lighting areas, generate corresponding lighting adjustment information according to the abnormal lighting areas and send the lighting adjustment information to the control module.

[0007] As a further improvement of the present invention, the illumination data is analyzed, and the specific analysis steps are as follows: By identifying the illumination data, the illumination environment data, the light source array data and the optical characteristic data are obtained; the illumination environment data includes the ambient light intensity and the illumination color, the illumination reference intensity preset for each sub-region is obtained, the ambient light intensity of each sub-region is calculated to be different from the illumination reference intensity to obtain the intensity difference, and it is recorded as AD; the illumination color of each sub-region is compared with the set standard reference color to obtain the color difference corresponding to each sub-region, the color difference is compared with the set color difference threshold, when the color difference is greater than the set color difference threshold, the color difference threshold is subtracted from the color difference to calculate the color difference influence value, and it is recorded as SC; Normalize the intensity difference and the color difference influence value and take their values, and use the formula The environmental impact value HJ is calculated; where u1 and u2 are preset weight factors, and their values ​​are 1.67 and 1.21 respectively.

[0008] The light source array data includes light source distribution, illumination angle and illumination position. The light source distribution is divided into a plurality of light source blocks of uniform size. The light source density of each light source block is obtained. A pre-set light source density interval is obtained. The light source density of each light source block is matched with the light source density interval. The light source blocks exceeding the light source density interval are marked as abnormal blocks. The number of abnormal blocks is counted and recorded as the number of abnormal blocks. The ratio of the number of abnormal blocks to the total number of light source blocks is calculated to obtain the abnormal block ratio YQ. Obtain the illumination angle of each light source block for multiple unit times, obtain a pre-set reference angle value, overlap and compare the illumination angle corresponding to each unit time with the reference angle value to obtain an angle deviation value; when the angle deviation value is greater than a set threshold, mark the corresponding angle deviation value as an instability angle value; add and sum the instability angle values ​​to obtain the instability impact value SW.

[0009] Get multiple set lighting reference points, compare the lighting position with multiple lighting reference points to get the lighting deviation points, count the number of lighting deviation points and record them as the number of deviation points, compare the number of deviation points with the set deviation point threshold, when the number of lighting deviation points is greater than the deviation point threshold, calculate the difference between the number of lighting deviation points and the deviation point threshold to get the deviation impact value PL.

[0010] The abnormal block ratio, instability impact value and deviation impact value are normalized and their values ​​are taken according to the formula The array influence value ZL is obtained; wherein ΔSW and ΔPL represent the instability influence allowable value and the deviation influence allowable value respectively; h1, h2 and h3 are all preset weight factors, and the values ​​are 2.19, 1.99 and 1.27 respectively.

[0011] The optical characteristic data are identified to obtain the reflection data and transmission data of the produced glass; the corresponding reflectivity is obtained according to the reflection data of the produced glass, the reflectivity is divided into multiple reflectivity intervals, each reflectivity interval corresponds to a reflection influence value, and the reflectivity corresponding to the produced glass is matched with the multiple reflectivity intervals to obtain the corresponding reflection influence value; the corresponding transmittance is obtained according to the transmission data of the produced glass, and the transmission influence value corresponding to the transmittance is obtained similarly; the reflection influence value and the transmission influence value are calculated and summed to obtain the optical influence value.

[0012] A quadrilateral is constructed with the values ​​of the environmental impact value and the array impact value as the long and short sides of the quadrilateral. A straight line perpendicular to the quadrilateral is drawn with the centroid of the quadrilateral as the starting point. The length of the straight line is equal to the value of the optical impact value. A quadrangular pyramid is constructed with the quadrilateral and the straight line. The volume of the quadrangular pyramid is calculated and the value of the volume is marked as the illumination state value. When the illumination state value is greater than the set threshold, the corresponding sub-area is marked as an abnormal illumination area.

[0013] The image processing module is used to obtain product image data and perform feature extraction on the product image data to obtain glass size data, surface quality information and cross-sectional image data, analyze the glass size data, surface quality information and cross-sectional image data to obtain glass quality information, and send the glass quality information to the feedback report module.

[0014] As a further improvement of the present invention, the glass size data, surface quality information and cross-sectional image data are analyzed, and the specific analysis steps are as follows: A1: Divide the produced glass into multiple glass blocks, obtain the glass size data of each glass block, and obtain the thickness data, length and width data and flatness data of the produced glass by identifying the glass size data; A11: Obtain the standard thickness of the produced glass, calculate the difference between the thickness data of each glass block and the standard thickness to obtain the thickness difference, obtain a preset thickness difference threshold, and when the thickness difference is greater than the thickness difference threshold, mark the corresponding thickness difference as a thickness abnormal value, and add up the thickness abnormal values ​​of each glass block to obtain the thickness impact value.

[0015] A12: Based on the length and width data, the length and width values ​​of the produced glass are obtained, the standard length of the produced glass is obtained, the length value corresponding to the produced glass is subtracted from the standard length to obtain the length difference, and a preset length difference threshold is obtained. When the length difference is greater than the length difference threshold, the length impact value is calculated by subtracting the length difference threshold from the length difference. Similarly, the width impact value corresponding to the width value is obtained, and the length impact value and the width impact value are calculated and summed to obtain the length and width impact value.

[0016] A13: The flatness of each glass block is obtained according to the flatness data of each glass block, and a preset flatness threshold is obtained. The flatness of each glass block is compared with the flatness threshold. When the flatness is greater than the flatness threshold, the corresponding glass section is marked as a flatness abnormal block, the number of flatness abnormal blocks is counted and recorded as the flatness abnormality number, and the flatness abnormality ratio is calculated by the ratio of the flatness abnormality number to the total number of glass blocks; the size status value is calculated by multiplying the thickness influence value, the length and width influence value and the flatness abnormality ratio.

[0017] A2: Obtain the surface quality information of each glass block corresponding to the produced glass, and obtain image information and roughness through recognition based on the surface quality information; A21: Identify the image information to obtain the defect features of each glass block, obtain the area of ​​the defect features, calculate and sum the defect feature areas of the glass blocks to obtain the block defect area, and then calculate and sum the block defect areas of each glass block to obtain the total defect area; defect features include bubbles, stones, scratches, stains, etc. The presence of defects will affect the appearance and performance of the glass and reduce product quality.

[0018] A22: The roughness corresponding to the produced glass is set to multiple roughness intervals, each roughness interval corresponds to a roughness impact value, and the roughness corresponding to the produced glass is matched with the multiple roughness intervals to obtain the corresponding roughness impact value.

[0019] A23: Construct an ellipse with the total defect area and the roughness influence value as the major axis and minor axis of the ellipse, calculate the area of ​​the ellipse and mark the area value as the surface state value.

[0020] A3: Obtain cross-sectional image data of the produced glass, obtain a cross-sectional texture image by identifying the cross-sectional image data, obtain a pre-set standard texture image, overlap and compare the cross-sectional texture image of the produced glass with the standard texture image to obtain a texture difference, compare the texture difference with a set texture difference threshold, and when the texture difference is greater than the texture difference threshold, mark the corresponding texture difference as a cross-sectional state value.

[0021] A4: Take the dimensional state value and the surface state value as the two right-angled sides of the right triangle, take the midpoint of the hypotenuse of the right triangle as the starting point, draw a straight line perpendicular to the hypotenuse, the length of the straight line is equal to the cross-sectional state value, connect the two ends of the hypotenuse with the ends of the straight line to form a quadrilateral, calculate the area of ​​the quadrilateral and mark the area of ​​the quadrilateral as the glass state value; when the glass state value is greater than the set threshold, the corresponding glass quality information is generated as quality abnormality.

[0022] The control module is used to receive the lighting adjustment information and generate adjustment instructions according to the lighting adjustment information, adjust the parameters according to the adjustment instructions, and obtain the lighting status information after the parameters are adjusted, compare the lighting status information before and after the adjustment to obtain the control effect, and send the control effect to the feedback report module.

[0023] The feedback report module is used to receive glass quality information and control effects. When the glass quality information is abnormal, a glass production abnormality report is generated; when the control effect is deteriorating, a control failure report is generated.

[0024] The data storage end is used to store product image data and lighting data, and to store glass production abnormality reports and control failure reports.

[0025] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects: The present invention obtains illumination status information by analyzing illumination data of each sub-area of ​​the illumination area, obtains illumination abnormality area according to the illumination information, and obtains corresponding illumination adjustment information according to the illumination abnormality area. The illumination information of the abnormal area can be adjusted through the illumination adjustment information to avoid the illumination abnormality reducing the effect of glass quality detection; the glass quality information is obtained by analyzing product image data, a glass production abnormality report is obtained according to the glass quality information, and a control failure report is generated according to the control effect, so as to facilitate the viewing of abnormal information. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. The following drawings are not intentionally scaled to the actual size, and the focus is on illustrating the main purpose of the present application.

[0027] Figure 1 It is a principle block diagram of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 In an embodiment of the present invention, an embodiment of a glass production quality inspection system based on machine vision includes: an image acquisition module, a lighting optimization analysis module, an image processing module, a control module, a feedback report module and a data storage terminal.

[0030] The image acquisition module collects light data through the light sensor, and uses a high-resolution industrial camera to capture the product image of the glass production line in real time to obtain product image data, and sends the product image data and light data to the data storage end; The lighting data includes lighting environment data, light source array data and optical characteristic data; the product image data includes glass size data, surface quality information and cross-sectional image data.

[0031] The lighting optimization analysis module divides the lighting area into multiple sub-areas, obtains the lighting data of each sub-area and analyzes the lighting data to obtain lighting status information, identifies the lighting status information to obtain the lighting abnormality area, generates corresponding lighting adjustment information according to the lighting abnormality area, and sends the lighting adjustment information to the control module; this module divides the lighting area into multiple sub-areas, analyzes the lighting data of each sub-area, identifies the lighting abnormality area, generates corresponding lighting adjustment information according to the abnormal area, and sends it to the control module; by adjusting the lighting, the image quality can be significantly improved and the negative impact of lighting abnormality on the glass quality detection effect can be avoided.

[0032] Analyze the illumination data. The specific analysis steps are as follows: By identifying the illumination data, the illumination environment data, the light source array data and the optical characteristic data are obtained; the illumination environment data includes the ambient light intensity and the illumination color, the illumination reference intensity set in advance for each sub-area is obtained, and the difference between the ambient light intensity of each sub-area and the illumination reference intensity is calculated to obtain the intensity difference; the illumination color of each sub-area is overlapped and compared with the set standard reference color to obtain the color difference corresponding to each sub-area, and the color difference is compared with the set color difference threshold. When the color difference is greater than the set color difference threshold, the color difference impact value is calculated by subtracting the color difference threshold from the color difference; the system analyzes the ambient light intensity and illumination color of each sub-area, and compares them with the set illumination reference intensity and the standard reference color to calculate the intensity difference and the color difference impact value; it can accurately identify abnormal areas in the illumination environment, such as excessive or weak illumination, color deviation, etc., and provide precise positioning for subsequent illumination adjustment.

[0033] Normalize the intensity difference and the color difference influence value and take their values, and use the formula The environmental impact value HJ is calculated, where AD and SC represent the intensity difference and color difference impact values ​​respectively; u1 and u2 are both preset weight factors, with values ​​of 1.67 and 1.21 respectively.

[0034] The light source array data includes light source distribution, lighting angle and lighting position. The light source distribution is divided into multiple light source blocks of uniform size, the light source density of each light source block is obtained, the pre-set light source density range is obtained, the light source density of each light source block is matched with the light source density range, the light source blocks exceeding the light source density range are marked as abnormal blocks, the number of abnormal blocks is counted and recorded as the number of abnormal blocks, and the ratio of the number of abnormal blocks to the total number of light source blocks is calculated to obtain the abnormal block ratio; by calculating the abnormal block ratio, the system can accurately identify the abnormal areas in the lighting distribution, provide precise targets for subsequent lighting adjustments, and make targeted adjustments to the abnormal blocks, which can significantly improve the uniformity of the entire lighting area, thereby ensuring the accuracy and stability of image acquisition during glass quality inspection.

[0035] Obtain the illumination angle of each light source block for multiple unit times, obtain the pre-set reference angle value, overlap and compare the illumination angle corresponding to each unit time with the reference angle value to obtain the angle deviation value. When the angle deviation value is greater than the set threshold, mark the corresponding angle deviation value as the instability angle value, and add and sum the instability angle values ​​to obtain the instability impact value. By calculating the instability impact value, the system can monitor the changes in illumination angle in real time, promptly discover and solve the problem of unstable illumination, and ensure the image acquisition quality during the glass quality inspection process. The stability of the illumination angle is crucial to avoid image blur and ghosting. By adjusting the illumination angle, the system can significantly reduce image quality problems caused by unstable illumination, thereby improving the accuracy of glass quality inspection.

[0036] Get multiple set lighting reference points, overlap and compare the lighting position with the multiple lighting reference points to get the lighting deviation points, count the number of lighting deviation points and record them as the number of deviation points, compare the number of deviation points with the set deviation point threshold, and when the number of lighting deviation points is greater than the deviation point threshold, calculate the difference between the number of lighting deviation points and the deviation point threshold to get the deviation impact value; accurate lighting position can ensure that the camera can capture clear and complete images of glass products, thereby improving the efficiency of image acquisition.

[0037] The abnormal block ratio, instability impact value and deviation impact value are normalized and their values ​​are taken according to the formula The array influence value ZL is obtained; among them, YQ, SW and PL represent the abnormal block proportion, instability influence value and deviation influence value respectively, ΔSW and ΔPL represent the instability influence allowable value and deviation influence allowable value respectively; h1, h2 and h3 are all preset weight factors, and their values ​​are 2.19, 1.99 and 1.27 respectively.

[0038] The optical characteristic data are identified to obtain the reflection data and transmission data of the produced glass; the corresponding reflectivity is obtained according to the reflection data of the produced glass, the reflectivity is divided into multiple reflectivity intervals, each reflectivity interval corresponds to a reflection influence value, and the reflectivity corresponding to the produced glass is matched with the multiple reflectivity intervals to obtain the corresponding reflection influence value; the corresponding transmittance is obtained according to the transmission data of the produced glass, and the transmission influence value corresponding to the transmittance is obtained similarly; the reflection influence value and the transmission influence value are calculated and summed to obtain the optical influence value.

[0039] A quadrilateral is constructed with the values ​​of the environmental impact value and the array impact value as the long and short sides of the quadrilateral. A straight line perpendicular to the quadrilateral is drawn with the centroid of the quadrilateral as the starting point. The length of the straight line is equal to the value of the optical impact value. A quadrangular pyramid is constructed with the quadrilateral and the straight line. The volume of the quadrangular pyramid is calculated and the value of the volume is marked as the illumination state value. When the illumination state value is greater than the set threshold, the corresponding sub-area is marked as an abnormal illumination area.

[0040] The image processing module acquires product image data and performs feature extraction on the product image data to obtain glass size data, surface quality information and cross-sectional image data, analyzes the glass size data, surface quality information and cross-sectional image data to obtain glass quality information, and sends the glass quality information to the feedback report module; the quality status of the glass is obtained by analyzing the quality-related data of the glass, thereby achieving the purpose of quality inspection of glass production.

[0041] The glass size data, surface quality information and cross-sectional image data are analyzed. The specific analysis steps are as follows: A1: Divide the produced glass into multiple glass blocks, obtain the glass size data of each glass block, and obtain the thickness data, length and width data and flatness data of the produced glass by identifying the glass size data; A11: Obtain the standard thickness of the produced glass, calculate the difference between the thickness data of each glass block and the standard thickness to obtain the thickness difference, obtain a preset thickness difference threshold, and when the thickness difference is greater than the thickness difference threshold, mark the corresponding thickness difference as a thickness abnormal value, and add up the thickness abnormal values ​​of each glass block to obtain the thickness impact value.

[0042] A12: Based on the length and width data, the length and width values ​​of the produced glass are obtained, the standard length of the produced glass is obtained, the length value corresponding to the produced glass is subtracted from the standard length to obtain the length difference, and a preset length difference threshold is obtained. When the length difference is greater than the length difference threshold, the length impact value is calculated by subtracting the length difference threshold from the length difference. Similarly, the width impact value corresponding to the width value is obtained, and the length impact value and the width impact value are calculated and summed to obtain the length and width impact value.

[0043] A13: The flatness of each glass block is obtained according to the flatness data of each glass block, and a preset flatness threshold is obtained. The flatness of each glass block is compared with the flatness threshold. When the flatness is greater than the flatness threshold, the corresponding glass section is marked as a flatness abnormal block, the number of flatness abnormal blocks is counted and recorded as the flatness abnormality number, and the flatness abnormality ratio is calculated by the ratio of the flatness abnormality number to the total number of glass blocks; the size status value is calculated by multiplying the thickness influence value, the length and width influence value and the flatness abnormality ratio.

[0044] A2: Obtain the surface quality information of each glass block corresponding to the produced glass, and obtain image information and roughness through recognition based on the surface quality information; A21: Identify the image information to obtain the defect features of each glass block, obtain the area of ​​the defect features, calculate and sum the defect feature areas of the glass blocks to obtain the block defect area, and then calculate and sum the block defect areas of each glass block to obtain the total defect area; defect features include bubbles, stones, scratches, stains, etc. The presence of defects will affect the appearance and performance of the glass and reduce product quality.

[0045] A22: The roughness corresponding to the produced glass is set to multiple roughness intervals, each roughness interval corresponds to a roughness impact value, and the roughness corresponding to the produced glass is matched with the multiple roughness intervals to obtain the corresponding roughness impact value.

[0046] A23: Construct an ellipse with the total defect area and the roughness influence value as the major axis and minor axis of the ellipse, calculate the area of ​​the ellipse and mark the area value as the surface state value.

[0047] A3: Obtain cross-sectional image data of the produced glass, obtain a cross-sectional texture image by identifying the cross-sectional image data, obtain a pre-set standard texture image, overlap and compare the cross-sectional texture image of the produced glass with the standard texture image to obtain a texture difference, compare the texture difference with a set texture difference threshold, and when the texture difference is greater than the texture difference threshold, mark the corresponding texture difference as a cross-sectional state value.

[0048] A4: Take the dimensional state value and the surface state value as the two right-angled sides of the right triangle, take the midpoint of the hypotenuse of the right triangle as the starting point, draw a straight line perpendicular to the hypotenuse, the length of the straight line is equal to the cross-sectional state value, connect the two ends of the hypotenuse with the ends of the straight line to form a quadrilateral, calculate the area of ​​the quadrilateral and mark the area of ​​the quadrilateral as the glass state value; when the glass state value is greater than the set threshold, the corresponding glass quality information is generated as quality abnormality.

[0049] The control module receives the light adjustment information and generates an adjustment instruction according to the light adjustment information, adjusts the parameters according to the adjustment instruction, obtains the light state information after the parameter adjustment, compares the light state information before and after the adjustment to obtain the control effect, and sends the control effect to the feedback report module; The control module can perform corresponding lighting adjustments on different blocks of the glass according to the adjustment instructions. The intelligent stepless dimming zone lamp can be used with the stepless dimming function to smoothly change the brightness of the lamp to facilitate the use of more needs, and the light-emitting blocks of the corresponding lamps can be zoned and controlled. Each light-emitting block can be independently adjusted in brightness, and the lighting effect can be controlled more finely, so that in a fixed lighting space, a sense of hierarchy can be created or a certain area can be highlighted by adjusting the brightness of different blocks. The system can flexibly adjust the light source according to actual conditions, ensuring the stability and consistency of glass quality detection.

[0050] The feedback report module receives glass quality information and control effects. When the glass quality information is abnormal, a glass production abnormality report is generated; when the control effect is deteriorating, a control failure report is generated.

[0051] The data storage end is used to store the product image data and lighting data sent by the image acquisition module and the lighting optimization analysis module; at the same time, it also stores the glass production abnormality report and control failure report generated by the feedback report module; the data storage end provides the system with persistent data storage capabilities.

[0052] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A glass production quality inspection system based on machine vision, including an image acquisition module, a feedback report module and a data storage terminal, characterized in that: Also includes: The lighting optimization analysis module is used to divide the lighting area into multiple sub-areas, obtain the lighting data of each sub-area and analyze the lighting data to obtain lighting status information, identify the lighting status information to obtain abnormal lighting areas, generate corresponding lighting adjustment information according to the abnormal lighting areas and send the lighting adjustment information to the control module; The image processing module is used to obtain product image data and perform feature extraction on the product image data to obtain glass size data, surface quality information and cross-sectional image data, analyze the glass size data, surface quality information and cross-sectional image data to obtain glass quality information, and send the glass quality information to the feedback report module; The control module is used to receive lighting adjustment information and generate adjustment instructions according to the lighting adjustment information, adjust parameters according to the adjustment instructions, and obtain lighting status information after the parameters are adjusted, compare the lighting status information before and after the adjustment to obtain the control effect, and send the control effect to the feedback report module.

2. The glass production quality inspection system based on machine vision according to claim 1, characterized in that: The image acquisition module collects light data through the light sensor, and uses a high-resolution industrial camera to capture product images of the glass production line in real time to obtain product image data, and sends the product image data and light data to the data storage end; The lighting data includes lighting environment data, light source array data and optical characteristic data; the product image data includes glass size data, surface quality information and cross-sectional image data.

3. The glass production quality inspection system based on machine vision according to claim 1, characterized in that: The illumination data is analyzed, and the specific analysis steps are as follows: The illumination environment data, the light source array data and the optical characteristic data are obtained by identifying the illumination data; the illumination environment data includes the ambient light intensity and the illumination color, the illumination reference intensity preset for each sub-region is obtained, and the ambient light intensity of each sub-region is calculated to obtain the intensity difference value by performing a difference calculation between the illumination reference intensity and the illumination color of each sub-region; the illumination color of each sub-region is compared with the set standard reference color to obtain the color difference value corresponding to each sub-region, and the color difference value is compared with the set color difference threshold value; when the color difference value is greater than the set color difference threshold value, the color difference value is subtracted from the color difference threshold value to calculate the color difference impact value; the intensity difference value and the color difference impact value are calculated comprehensively to obtain the environment impact value; The light source array data includes light source distribution, illumination angle and illumination position. The light source distribution is divided into a plurality of light source blocks of uniform size. The light source density of each light source block is obtained. A pre-set light source density interval is obtained. The light source density of each light source block is matched with the light source density interval. The light source blocks exceeding the light source density interval are marked as abnormal blocks. The number of abnormal blocks is counted and recorded as the number of abnormal blocks. The ratio of the number of abnormal blocks to the total number of light source blocks is calculated to obtain the abnormal block ratio. Obtain the illumination angle of each light source block for multiple unit times, obtain a preset reference angle value, overlap and compare the illumination angle corresponding to each unit time with the reference angle value to obtain an angle deviation value, and when the angle deviation value is greater than a set threshold, mark the corresponding angle deviation value as an instability angle value, and add and sum the instability angle values ​​to obtain an instability impact value; Get multiple set illumination reference points, compare the illumination position with the multiple illumination reference points to obtain illumination deviation points, count the number of illumination deviation points and record them as deviation point counts, compare the deviation point counts with the set deviation point threshold, and when the illumination deviation point count is greater than the deviation point threshold, calculate the difference between the illumination deviation point count and the deviation point threshold to obtain the deviation impact value; The array impact value is obtained by comprehensively calculating the abnormal block ratio, instability impact value and deviation impact value; Reflection data and transmission data of the produced glass are obtained by identifying the optical characteristic data; the corresponding reflectivity is obtained according to the reflection data of the produced glass, the reflectivity is divided into a plurality of reflectivity intervals, each reflectivity interval corresponds to a reflection influence value, and the reflectivity corresponding to the produced glass is matched with the plurality of reflectivity intervals to obtain the corresponding reflection influence value; the corresponding transmittance is obtained according to the transmission data of the produced glass, and the transmission influence value corresponding to the transmittance is obtained similarly; the reflection influence value and the transmission influence value are calculated and summed to obtain the optical influence value; A quadrilateral is constructed with the values ​​of the environmental impact value and the array impact value as the long and short sides of the quadrilateral, a straight line is drawn perpendicular to the quadrilateral with the center of gravity of the quadrilateral as the starting point, and the length of the straight line is equal to the value of the optical impact value, and then a quadrangular pyramid is constructed with the quadrilateral and the straight line, the volume of the quadrangular pyramid is calculated and the value of the volume is marked as the illumination state value; When the illumination state value is greater than the set threshold, the corresponding sub-region is marked as an abnormal illumination region.

4. The glass production quality inspection system based on machine vision according to claim 1, characterized in that: The glass size data, surface quality information and cross-sectional image data are analyzed, and the specific analysis steps are as follows: A1: Divide the produced glass into multiple glass blocks, obtain the glass size data of each glass block, and obtain the thickness data, length and width data and flatness data of the produced glass by identifying the glass size data; A11: Obtain the standard thickness of the produced glass, calculate the difference between the thickness data of each glass block and the standard thickness to obtain the thickness difference, obtain a preset thickness difference threshold, and when the thickness difference is greater than the thickness difference threshold, mark the corresponding thickness difference as a thickness abnormal value, and add up the thickness abnormal values ​​of each glass block to obtain the thickness impact value; A12: Based on the length and width data, the length and width of the produced glass are obtained, the standard length of the produced glass is obtained, the length value corresponding to the produced glass is calculated to be different from the standard length to obtain the length difference, and a preset length difference threshold is obtained. When the length difference is greater than the length difference threshold, the length difference threshold is subtracted from the length difference to obtain the length influence value. Similarly, the width influence value corresponding to the width value is obtained, and the length influence value and the width influence value are calculated and summed to obtain the length and width influence value. A13: Obtain the flatness of each glass block according to the flatness data of each glass block, obtain a preset flatness threshold, compare the flatness of each glass block with the flatness threshold, and when the flatness is greater than the flatness threshold, mark the corresponding glass section as a flatness abnormal block, count the number of flatness abnormal blocks and record it as the flatness abnormal number, calculate the ratio of the flatness abnormal number to the total number of glass blocks, and calculate the size status value by multiplying the thickness influence value, the length and width influence value, and the flatness abnormality ratio; A2: Obtain the surface quality information of each glass block corresponding to the produced glass, and obtain image information and roughness through recognition based on the surface quality information; A21: Identify the image information to obtain defect features of each glass block, obtain the area of ​​the defect features, calculate and sum the defect feature areas of the glass blocks to obtain the block defect area, and then calculate and sum the block defect areas of each glass block to obtain the total defect area; A22: setting a plurality of roughness intervals for the roughness corresponding to the produced glass, each roughness interval corresponding to a roughness impact value, and matching the roughness corresponding to the produced glass with the plurality of roughness intervals to obtain the corresponding roughness impact value; A23: Construct an ellipse with the total defect area and the roughness influence value as the major axis and minor axis of the ellipse, calculate the area of ​​the ellipse and mark the area value as the surface state value; A3: Obtain cross-sectional image data of the produced glass, obtain a cross-sectional texture image by identifying the cross-sectional image data, obtain a preset standard texture image, overlap and compare the cross-sectional texture image of the produced glass with the standard texture image to obtain a texture difference, compare the texture difference with a preset texture difference threshold, and when the texture difference is greater than the texture difference threshold, mark the corresponding texture difference as a cross-sectional state value; A4: Take the dimensional state value and the surface state value as the two right-angled sides of the right triangle, take the midpoint of the hypotenuse of the right triangle as the starting point, draw a straight line perpendicular to the hypotenuse, the length of the straight line is equal to the cross-sectional state value, connect the two ends of the hypotenuse with the ends of the straight line to form a quadrilateral, calculate the area of ​​the quadrilateral and mark the area of ​​the quadrilateral as the glass state value; when the glass state value is greater than the set threshold, the corresponding glass quality information is generated as quality abnormality.

5. The glass production quality inspection system based on machine vision according to claim 1, characterized in that: The feedback report module is used to receive glass quality information and control effects. When the glass quality information is abnormal, a glass production abnormality report is generated; when the control effect is deteriorating, a control failure report is generated.

Images