A digital detection system and method for glass edge grinding based on 3D visual detection

Abstract

The present invention discloses a digital detection system and method for glass edge grinding based on 3D visual detection, which belongs to the technical field of glass processing. The system includes a measurement data acquisition module, a 3D model management module, a surface roughness analysis module, a digital evaluation module and an intelligent control module; the measurement data acquisition module is used for collecting image data formed by laser lines on the glass surface by a camera after the glass edge grinding equipment performs edge grinding on the glass; the 3D model management module is used for analyzing the collected image data, establishing a 3D space model, and determining the spatial shape of the glass surface based on the triangulation principle; the surface roughness analysis module is used for determining the surface roughness of the glass after edge grinding; the digital evaluation module is used for judging whether edge grinding operation is required based on comprehensive evaluation of the surface roughness of the glass after current edge grinding; the intelligent control module is used for controlling the glass edge grinding equipment to perform edge grinding on the glass that needs to be edged.

Description

Technical Field

[0001] The present invention relates to the technical field of glass processing, and in particular to a glass edge grinding digital detection system and method based on 3D visual detection. Background Art

[0002] 3D visual inspection is a method of detecting and identifying objects using computer vision technology combined with 3D imaging equipment. It can obtain three-dimensional information of objects, including shape, size, position, etc., and is widely used in fields such as industrial automation. Digital inspection of glass edge grinding through 3D visual inspection can effectively detect tiny defects that are difficult for the human eye to distinguish, ensuring production efficiency. In the existing technology, the surface roughness of the glass edge grinding is often used to judge whether the current edge grinding is qualified. If the surface roughness value is within the specified tolerance range, the edge grinding is considered to be qualified. However, this judgment method is difficult to comprehensively evaluate the edge grinding quality, and the edge grinding effects of glass of different models and qualities under the same process will also be different. Summary of the invention

[0003] The object of the present invention is to provide a glass edge grinding digital detection system and method based on 3D visual detection to solve the problems raised in the above background technology.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions: a digital detection method for glass edge grinding based on 3D visual detection, the method comprising the following steps:

[0005] S10, after the glass edge grinding equipment performs edge grinding on the glass, a camera is used to collect image data formed by the laser line on the glass surface;

[0006] S20, analyzing the collected image data, establishing a 3D spatial model, and determining the spatial shape of the glass surface based on the triangulation principle; and determining the surface roughness of the glass after edge grinding according to the spatial shape of the glass surface;

[0007] S30, based on the image data of the historical glass after edge grinding, comprehensively evaluate the surface roughness of the glass after the current edge grinding, and determine whether the edge grinding operation is required; if the edge grinding operation is required, execute step S40; if the edge grinding operation is not required, execute step S50;

[0008] S40, controlling the glass edge grinding equipment to perform re-edge grinding on the glass that needs to be edged, and repeating step S10;

[0009] S50, glass products are qualified and edge grinding is completed.

[0010] Furthermore, the glass edge grinding device is equipped with a calibration laser and a camera; the calibration laser is used to emit a laser line to the glass surface and calibrate the relative position with the camera; the camera is used to capture image data formed by the laser line on the glass surface; when the glass edge grinding device performs edge grinding on the glass, the calibration laser is controlled to calibrate the relative position with the camera, and a laser line is emitted to the glass surface, and the image data is collected using the camera.

[0011] Furthermore, the method steps of step S20 are as follows:

[0012] S201, establish a 3D spatial model about the x-axis, y-axis and z-axis on the edged glass surface, determine the focal length f of the camera and the coordinates of the center point of the camera imaging plane (x0, y0), based on the principle of triangulation, according to the focal length and center point coordinates of the camera, respectively determine the spatial coordinates (X, Y, Z) of the edge point of the glass surface, according to the calculation formula:

[0013]

[0014] Where M represents the tangent value of the angle of the laser line in the camera's field of view; Δx represents the change in the pixel position of the laser line on the image; X0 and Y0 represent the pixel coordinates of the center of the laser line on the image respectively;

[0015] By calculating the coordinate set of the space points on the edge-grinding glass surface, the roughness of the glass surface after edge-grinding is calculated;

[0016] S202, according to the calculated spatial coordinates of the edge points of the glass surface, a set of spatial coordinates of the edge points of the glass surface is obtained, and the height function F(x) of the glass surface is determined according to the set, and the sample length L is determined; according to the selected sample length and the height function of the glass surface, the roughness of the glass surface after edge grinding is determined according to the calculation formula:

[0017]

[0018] Where C represents the calculated surface roughness of the glass after edge grinding; represents the average value of F(x); F(x) represents the surface height change of the glass along the measuring direction x; the sample length L represents the characteristics of the entire glass surface;

[0019] Calculate the roughness of ground glass surfaces for a first evaluation of glass products.

[0020] Furthermore, the method steps of step S30 are as follows:

[0021] S301, gridding the collected image data, dividing the image into m grids; determining the maximum threshold of the surface roughness of the glass edge grinding according to the type and use of the glass

[0022] Where N represents the number of divided grids;

[0023] S302, based on the qualified image data of the historical glass after edge grinding, respectively determine the surface roughness of the glass after edge grinding under each grid to obtain a set P; P = {P1, P2, ..., P n};P i ={p i1 、p i2 , ..., p im};

[0024] Among them, P i represents the image data of the i-th historical glass after edge grinding; p ij represents the glass surface roughness in the jth grid in the image data after the i-th historical glass edge grinding; i = 1, 2, ..., n; j = 1, 2, ..., m;

[0025] p ij The calculation formula is:

[0026]

[0027] Among them, L j represents the sample length selected in the jth grid; F ij (x) represents the glass surface height function determined by the jth grid in the image data after the i-th historical glass edge grinding; Indicates F ij The average value of (x);

[0028] S303, according to the set P, respectively calculate the variance value between the glass surface roughness in each grid in the image data, according to the calculation formula:

[0029]

[0030] Among them, Z i represents the calculated variance value of the i-th historical glass edge grinding image data; represents the average value of glass surface roughness in each grid in the i-th historical glass edge grinding image data;

[0031] Calculate the variance of the surface roughness of the ground glass to conduct a second evaluation of the glass product;

[0032] According to the calculated Z1, Z2, ..., Z n , get the set W;

[0033] S304, construct a BP neural network, divide the set W into a training set W1 and a validation set W2, train the network weights based on the training set W1 through a back propagation algorithm, and then use the validation set W2 to evaluate the network performance through cross-validation to determine the maximum threshold of the variance after the glass is edge-polished.

[0034] S305, determining the surface roughness C0 of the glass after the current edge grinding, and determining the variance value Z0 of the glass image data after the current edge grinding, and judging whether the edge grinding operation is required; when and When there is no need to perform edge grinding; when there is or When cutting, edge grinding is required.

[0035] Furthermore, the glass edge grinding equipment can digitally display the surface roughness of the glass after edge grinding and the calculated variance value; when the glass product is qualified and the edge grinding is completed, a production number of the corresponding glass product is generated, and the management personnel can view the surface roughness and the calculated variance value of the glass after production through the production number.

[0036] A digital glass edge grinding detection system based on 3D visual detection, the system includes a measurement data acquisition module, a 3D model management module, a surface roughness analysis module, a digital evaluation module and an intelligent control module;

[0037] The measurement data acquisition module is used to use a camera to collect image data formed by the laser line on the glass surface after the glass edge grinding equipment has processed the glass edge. The 3D model management module is used to analyze the collected image data, establish a 3D spatial model, and determine the spatial shape of the glass surface based on the triangulation principle. The surface roughness analysis module is used to determine the surface roughness of the glass after edge grinding according to the spatial shape of the glass surface. The digital evaluation module is used to comprehensively evaluate the surface roughness of the current glass after edge grinding based on the image data of the historical glass after edge grinding, and determine whether edge grinding operation is required. The intelligent control module is used to control the glass edge grinding equipment to perform edge grinding on the glass that needs to be edged.

[0038] Further, the measurement data acquisition module includes a calibration laser and a camera;

[0039] The calibration laser is used to emit a laser line to the glass surface and calibrate the relative position with the camera;

[0040] The camera is used to capture image data formed by the laser line on the glass surface;.

[0041] Further, the 3D model management module includes a model building unit, a data setting unit and a coordinate point determination unit;

[0042] The model building unit is used to build a 3D spatial model about the x-axis, y-axis and z-axis on the edge-grinding glass surface;

[0043] The data setting unit is used to determine the focal length of the camera and the coordinates of the center point of the camera imaging plane;

[0044] The coordinate point determination unit is used to determine the spatial coordinates of the edge points of the glass surface respectively according to the focal length and the center point coordinates of the camera based on the triangulation principle.

[0045] Further, the surface roughness analysis module includes a surface function determination unit and a surface roughness determination unit;

[0046] The surface function determination unit is used to obtain a set of spatial coordinates of the edge points of the glass surface according to the calculated spatial coordinates of the edge points of the glass surface, and determine the height function of the glass surface according to the set;

[0047] The surface roughness determination unit is used to determine the sample length, and determine the surface roughness of the glass after edge grinding according to the selected sample length and the height function of the glass surface.

[0048] Furthermore, the digital evaluation module includes a grid division unit, a historical data analysis unit, a variance value calculation unit, a threshold determination unit and an analysis and judgment unit;

[0049] The grid division unit is used to perform grid division on the collected image data;

[0050] The historical data analysis unit is used to determine the surface roughness of the glass after edge grinding under each grid based on the qualified image data of the historical glass after edge grinding;

[0051] The variance value calculation unit is used to respectively calculate the variance value between the glass surface roughness in each grid in the image data;

[0052] The threshold determination unit is used to determine the maximum threshold of surface roughness of the glass edge grinding according to the type and purpose of the glass; construct a BP neural network to determine the maximum threshold of variance of the glass after edge grinding;

[0053] The analysis and judgment unit is used to comprehensively evaluate the surface roughness of the glass after the current edge grinding, and judge whether the edge grinding operation is needed; if the edge grinding operation is needed, the glass edge grinding equipment is controlled to re-edge the glass that needs to be edged; if the edge grinding operation is not needed, the glass product is qualified and the edge grinding is completed.

[0054] Compared with the prior art, the beneficial effects achieved by the present invention are: providing a digital detection system and method for glass edge grinding based on 3D visual detection, by making a secondary judgment on the surface roughness of the glass edge grinding and calculating the variance value of the image data after the glass edge grinding, the accuracy of the glass edge grinding quality assessment is improved; at the same time, by digitally displaying the data analysis results, it can help managers better understand product quality and meet the needs of digital detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0056] Figure 1 It is a structural schematic diagram of a glass edge grinding digital detection system based on 3D visual detection in the present invention. DETAILED DESCRIPTION

[0057] 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.

[0058] See also Figure 1 The present invention provides a technical solution: a digital detection method for glass edge grinding based on 3D visual detection, the method comprising the following steps:

[0059] S10, after the glass edge grinding equipment performs edge grinding on the glass, a camera is used to collect image data formed by the laser line on the glass surface;

[0060] S20, analyzing the collected image data, establishing a 3D spatial model, and determining the spatial shape of the glass surface based on the triangulation principle; and determining the surface roughness of the glass after edge grinding according to the spatial shape of the glass surface;

[0061] S30, based on the image data of the historical glass after edge grinding, comprehensively evaluate the surface roughness of the glass after the current edge grinding, and determine whether the edge grinding operation is required; if the edge grinding operation is required, execute step S40; if the edge grinding operation is not required, execute step S50;

[0062] S40, controlling the glass edge grinding equipment to perform re-edge grinding on the glass that needs to be edged, and repeating step S10;

[0063] S50, glass products are qualified and edge grinding is completed.

[0064] Furthermore, the glass edge grinding device is equipped with a calibration laser and a camera; the calibration laser is used to emit a laser line to the glass surface and calibrate the relative position with the camera; the camera is used to capture image data formed by the laser line on the glass surface; when the glass edge grinding device performs edge grinding on the glass, the calibration laser is controlled to calibrate the relative position with the camera, and a laser line is emitted to the glass surface, and the image data is collected using the camera.

[0065] Furthermore, the method steps of step S20 are as follows:

[0066] S201, establish a 3D spatial model about the x-axis, y-axis and z-axis on the edged glass surface, determine the focal length f of the camera and the coordinates of the center point of the camera imaging plane (x0, y0), based on the principle of triangulation, according to the focal length and center point coordinates of the camera, respectively determine the spatial coordinates (X, Y, Z) of the edge point of the glass surface, according to the calculation formula:

[0067]

[0068] Where M represents the tangent value of the angle of the laser line in the camera's field of view; Δx represents the change in the pixel position of the laser line on the image; X0 and Y0 represent the pixel coordinates of the center of the laser line on the image respectively;

[0069] S202, according to the calculated spatial coordinates of the edge points of the glass surface, a set of spatial coordinates of the edge points of the glass surface is obtained, and the height function F(x) of the glass surface is determined according to the set, and the sample length L is determined; according to the selected sample length and the height function of the glass surface, the roughness of the glass surface after edge grinding is determined according to the calculation formula:

[0070]

[0071] Where C represents the calculated surface roughness of the glass after edge grinding; represents the average value of F(x); F(x) represents the surface height variation of the glass along the measuring direction x; and the sample length L represents the characteristics of the entire glass surface.

[0072] Furthermore, the method steps of step S30 are as follows:

[0073] S301, gridding the collected image data, dividing the image into m grids; determining the maximum threshold of the surface roughness of the glass edge grinding according to the type and use of the glass

[0074] Where N represents the number of divided grids;

[0075] S302, based on the qualified image data of the historical glass after edge grinding, respectively determine the surface roughness of the glass after edge grinding under each grid to obtain a set P; P = {P1, P2, ..., P n};P i ={p i1 、p i2 , ..., p im};

[0076] Among them, P i represents the image data of the i-th historical glass after edge grinding; p ij represents the glass surface roughness in the jth grid in the image data after the i-th historical glass edge grinding; i = 1, 2, ..., n; j = 1, 2, ..., m;

[0077] p ij The calculation formula is:

[0078]

[0079] Among them, L j represents the sample length selected in the jth grid; F ij (x) represents the glass surface height function determined by the jth grid in the image data after the i-th historical glass edge grinding; Indicates F ij The average value of (x);

[0080] S303, according to the set P, respectively calculate the variance value between the glass surface roughness in each grid in the image data, according to the calculation formula:

[0081]

[0082] Among them, Z i represents the calculated variance value of the i-th historical glass edge grinding image data; represents the average value of glass surface roughness in each grid in the i-th historical glass edge grinding image data;

[0083] According to the calculated Z1, Z2, ..., Z n , get the set W;

[0084] S304, construct a BP neural network, divide the set W into a training set W1 and a validation set W2, train the network weights based on the training set W1 through a back propagation algorithm, and then use the validation set W2 to evaluate the network performance through cross-validation to determine the maximum threshold of the variance after the glass is edge-polished.

[0085] S305, determining the surface roughness C0 of the glass after the current edge grinding, and determining the variance value Z0 of the glass image data after the current edge grinding, and judging whether the edge grinding operation is required; when and When there is no need to perform edge grinding; when there is or When cutting, edge grinding is required.

[0086] Furthermore, the glass edge grinding equipment can digitally display the surface roughness of the glass after edge grinding and the calculated variance value; when the glass product is qualified and the edge grinding is completed, a production number of the corresponding glass product is generated, and the management personnel can view the surface roughness and the calculated variance value of the glass after production through the production number.

[0087] A digital glass edge grinding detection system based on 3D visual detection, the system includes a measurement data acquisition module, a 3D model management module, a surface roughness analysis module, a digital evaluation module and an intelligent control module;

[0088] The measurement data acquisition module is used to use a camera to collect image data formed by the laser line on the glass surface after the glass edge grinding equipment has processed the glass edge. The 3D model management module is used to analyze the collected image data, establish a 3D spatial model, and determine the spatial shape of the glass surface based on the triangulation principle. The surface roughness analysis module is used to determine the surface roughness of the glass after edge grinding according to the spatial shape of the glass surface. The digital evaluation module is used to comprehensively evaluate the surface roughness of the current glass after edge grinding based on the image data of the historical glass after edge grinding, and determine whether edge grinding operation is required. The intelligent control module is used to control the glass edge grinding equipment to perform edge grinding on the glass that needs to be edged.

[0089] The measurement data acquisition module includes a calibration laser and a camera;

[0090] The calibration laser is used to emit a laser line to the glass surface and calibrate the relative position with the camera;

[0091] The camera is used to capture image data formed by the laser line on the glass surface;.

[0092] Further, the 3D model management module includes a model building unit, a data setting unit and a coordinate point determination unit;

[0093] The model building unit is used to build a 3D spatial model about the x-axis, y-axis and z-axis on the edge-grinding glass surface;

[0094] The data setting unit is used to determine the focal length of the camera and the coordinates of the center point of the camera imaging plane;

[0095] The coordinate point determination unit is used to determine the spatial coordinates of the edge points of the glass surface respectively according to the focal length and the center point coordinates of the camera based on the triangulation principle.

[0096] Further, the surface roughness analysis module includes a surface function determination unit and a surface roughness determination unit;

[0097] The surface function determination unit is used to obtain a set of spatial coordinates of the edge points of the glass surface according to the calculated spatial coordinates of the edge points of the glass surface, and determine the height function of the glass surface according to the set;

[0098] The surface roughness determination unit is used to determine the sample length, and determine the surface roughness of the glass after edge grinding according to the selected sample length and the height function of the glass surface.

[0099] Furthermore, the digital evaluation module includes a grid division unit, a historical data analysis unit, a variance value calculation unit, a threshold determination unit and an analysis and judgment unit;

[0100] The grid division unit is used to perform grid division on the collected image data;

[0101] The historical data analysis unit is used to determine the surface roughness of the glass after edge grinding under each grid based on the qualified image data of the historical glass after edge grinding;

[0102] The variance value calculation unit is used to respectively calculate the variance value between the glass surface roughness in each grid in the image data;

[0103] The threshold determination unit is used to determine the maximum threshold of surface roughness of the glass edge grinding according to the type and purpose of the glass; construct a BP neural network to determine the maximum threshold of variance of the glass after edge grinding;

[0104] The analysis and judgment unit is used to comprehensively evaluate the surface roughness of the glass after the current edge grinding, and judge whether the edge grinding operation is needed; if the edge grinding operation is needed, the glass edge grinding equipment is controlled to re-edge the glass that needs to be edged; if the edge grinding operation is not needed, the glass product is qualified and the edge grinding is completed.

[0105] In this embodiment:

[0106] The system is specifically a single-sided glass edge grinding digital detection system; in the system, a glass straight edge grinding machine is used to grind the glass product; the glass straight edge grinding machine is equipped with a calibration laser and a camera;

[0107] A 3D spatial model about the x-axis, y-axis and z-axis is established on the edge-grinding glass surface to determine the focal length f of the camera and the coordinates of the center point of the camera imaging plane (x0, y0).

[0108] After the glass straight edge grinding machine grinds the glass, the camera is used to collect the image data formed by the laser line on the glass surface. Based on the principle of triangulation, the spatial coordinates (X, Y, Z) of the edge points on the glass surface are determined according to the focal length and center point coordinates of the camera. According to the calculation formula:

[0109]

[0110] According to the calculated spatial coordinates of the edge points of the glass surface, the spatial coordinate set of the edge points of the glass surface is obtained, and the height function F(x) of the glass surface is determined according to the set, and the sample length L is determined; according to the selected sample length and the height function of the glass surface, the roughness of the glass surface after edge grinding is determined according to the calculation formula:

[0111]

[0112] The collected image data is gridded and the image is divided into 20 grids. According to the type and purpose of glass, the maximum threshold of surface roughness for qualified glass edge grinding is determined.

[0113] Based on the qualified image data of historical glass after edge grinding, the surface roughness of the glass after edge grinding under each grid is determined respectively, and the set P is obtained; P = {P1, P2, ..., P n};P i ={p i1 、p i2 , ..., p im};

[0114] p ij The calculation formula is:

[0115]

[0116] According to the set P, the variance values ​​of the glass surface roughness in each grid in the image data are calculated respectively, according to the calculation formula:

[0117]

[0118] According to the calculated Z1, Z2, ..., Z n , get the set W; construct a BP neural network, divide the set W into a training set W1 and a validation set W2, train the network weights based on the training set W1 through the back propagation algorithm, and then use the validation set W2 to evaluate the network performance through cross-validation to determine the maximum threshold of the variance after the glass is edged.

[0119] At a certain point:

[0120] Determine the surface roughness C0 of the glass after the current edge grinding, and determine the variance value Z0 of the glass image data after the current edge grinding; perform a comprehensive evaluation on the surface roughness of the glass after the current edge grinding, and at this time, Z0> If edge grinding is required, the glass edge grinding equipment is controlled to re-edge the glass that needs to be edged, and the edged glass is re-judged;

[0121] Among them, when the glass product is qualified and the edge grinding is completed, the production number of the corresponding glass product is generated. The management personnel can check the surface roughness and calculated variance value of the glass after production through the production number.

[0122] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0123] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A digital detection method for glass edge grinding based on 3D visual detection, characterized in that: The method comprises the following steps: S10, after the glass edge grinding equipment performs edge grinding on the glass, a camera is used to collect image data formed by the laser line on the glass surface; S20, analyzing the collected image data, establishing a 3D spatial model, and determining the spatial shape of the glass surface based on the triangulation principle; and determining the surface roughness of the glass after edge grinding according to the spatial shape of the glass surface; S30, based on the image data of the historical glass after edge grinding, comprehensively evaluate the surface roughness of the glass after the current edge grinding, and determine whether the edge grinding operation is required; if the edge grinding operation is required, execute step S40; if the edge grinding operation is not required, execute step S50; S40, controlling the glass edge grinding equipment to perform re-edge grinding on the glass that needs to be edged, and repeating step S10; S50, glass products are qualified and edge grinding is completed; The method steps of step S20 are: S201, establish a 3D spatial model about the x-axis, y-axis and z-axis on the edged glass surface, determine the focal length f of the camera and the coordinates of the center point of the camera imaging plane (x0, y0), based on the principle of triangulation, according to the focal length and center point coordinates of the camera, respectively determine the spatial coordinates (X, Y, Z) of the edge point of the glass surface, according to the calculation formula: Where M represents the tangent value of the angle of the laser line in the camera's field of view; △x represents the change in the pixel position of the laser line on the image; X0 and Y0 represent the pixel coordinates of the center of the laser line on the image respectively; S202, according to the calculated spatial coordinates of the edge points of the glass surface, a set of spatial coordinates of the edge points of the glass surface is obtained, and the height function F(x) of the glass surface is determined according to the set, and the sample length L is determined; according to the selected sample length and the height function of the glass surface, the roughness of the glass surface after edge grinding is determined according to the calculation formula: Where C represents the calculated surface roughness of the glass after edge grinding; represents the average value of F(x); F(x) represents the surface height change of the glass along the measuring direction x; the sample length L represents the characteristics of the entire glass surface; The method steps of step S30 are: S301, gridding the collected image data, dividing the image into m grids; determining the maximum threshold of the surface roughness of the glass edge grinding according to the type and use of the glass Where N represents the number of divided grids; S302, based on the qualified image data of the historical glass after edge grinding, respectively determine the surface roughness of the glass after edge grinding under each grid to obtain a set P; P = {P1, P2, ..., P n };P i ={p i1 、p i2 , ..., p im }; Among them, P i represents the image data of the i-th historical glass after edge grinding; p ij represents the glass surface roughness in the jth grid in the image data after the i-th historical glass edge grinding; i = 1, 2, ..., n; j = 1, 2, ..., m; p ij The calculation formula is: Among them, L j represents the sample length selected in the jth grid; F ij (x) represents the glass surface height function determined by the jth grid in the image data after the i-th historical glass edge grinding; Indicates F ij The average value of (x); S303, according to the set P, respectively calculate the variance value between the glass surface roughness in each grid in the image data, according to the calculation formula: Among them, Z i represents the calculated variance value of the i-th historical glass edge grinding image data; represents the average value of glass surface roughness in each grid in the i-th historical glass edge grinding image data; According to the calculated Z1, Z2, ..., Z n , get the set W; S304, construct a BP neural network, divide the set W into a training set W1 and a validation set W2, train the network weights based on the training set W1 through a back propagation algorithm, and then use the validation set W2 to evaluate the network performance through cross-validation to determine the maximum threshold of the variance after the glass is edge-polished. S305, determining the surface roughness C0 of the glass after the current edge grinding, and determining the variance value Z0 of the glass image data after the current edge grinding, and judging whether the edge grinding operation is required; and When there is no need to perform edge grinding; when there is or When cutting, edge grinding is required.

2. The glass edge grinding digital detection method based on 3D visual detection according to claim 1, characterized in that: The glass edge grinding device is equipped with a calibration laser and a camera; the calibration laser is used to emit a laser line to the glass surface and calibrate the relative position with the camera; the camera is used to capture the image data formed by the laser line on the glass surface; after the glass edge grinding device performs edge grinding on the glass, the calibration laser is controlled to calibrate the relative position with the camera, and a laser line is emitted to the glass surface, and the camera is used to collect image data.

3. The glass edge grinding digital detection method based on 3D visual detection according to claim 2 is characterized in that: The glass edge grinding equipment can digitally display the surface roughness of the glass after edge grinding and the calculated variance value; when the glass product is qualified and the edge grinding is completed, a production number of the corresponding glass product is generated, and the management personnel can view the surface roughness and the calculated variance value of the glass after production through the production number.

4. A digital inspection system for glass edge grinding based on 3D visual inspection, using the digital inspection method for glass edge grinding based on 3D visual inspection as claimed in claim 1, characterized in that: The system includes a measurement data acquisition module, a 3D model management module, a surface roughness analysis module, a digital evaluation module and an intelligent control module; The measurement data acquisition module is used to use a camera to collect image data formed by the laser line on the glass surface after the glass edge grinding equipment has processed the glass edge. The 3D model management module is used to analyze the collected image data, establish a 3D spatial model, and determine the spatial shape of the glass surface based on the triangulation principle. The surface roughness analysis module is used to determine the surface roughness of the glass after edge grinding according to the spatial shape of the glass surface. The digital evaluation module is used to comprehensively evaluate the surface roughness of the current glass after edge grinding based on the image data of the historical glass after edge grinding, and determine whether edge grinding operation is required. The intelligent control module is used to control the glass edge grinding equipment to perform edge grinding on the glass that needs to be edged.

5. The glass edge grinding digital detection system based on 3D visual detection according to claim 4 is characterized in that: The measurement data acquisition module includes a calibration laser and a camera; The calibration laser is used to emit a laser line to the glass surface and calibrate the relative position with the camera; The camera is used to capture image data formed by the laser line on the glass surface.

6. The glass edge grinding digital detection system based on 3D visual detection according to claim 5, characterized in that: The 3D model management module includes a model building unit, a data setting unit and a coordinate point determination unit; The model building unit is used to build a 3D spatial model about the x-axis, y-axis and z-axis on the edge-grinding glass surface; The data setting unit is used to determine the focal length of the camera and the coordinates of the center point of the camera imaging plane; The coordinate point determination unit is used to determine the spatial coordinates of the edge points of the glass surface respectively according to the focal length and the center point coordinates of the camera based on the triangulation principle.

7. The glass edge grinding digital detection system based on 3D visual detection according to claim 6 is characterized by: The surface roughness analysis module includes a surface function determination unit and a surface roughness determination unit; The surface function determination unit is used to obtain a set of spatial coordinates of the edge points of the glass surface according to the calculated spatial coordinates of the edge points of the glass surface, and determine the height function of the glass surface according to the set; The surface roughness determination unit is used to determine the sample length, and determine the surface roughness of the glass after edge grinding according to the selected sample length and the height function of the glass surface.

8. The glass edge grinding digital detection system based on 3D visual detection according to claim 7 is characterized in that: The digital evaluation module includes a grid division unit, a historical data analysis unit, a variance value calculation unit, a threshold determination unit and an analysis and judgment unit; The grid division unit is used to perform grid division on the collected image data; The historical data analysis unit is used to determine the surface roughness of the glass after edge grinding under each grid based on the qualified image data of the historical glass after edge grinding; The variance value calculation unit is used to respectively calculate the variance value between the glass surface roughness in each grid in the image data; The threshold determination unit is used to determine the maximum threshold of surface roughness for qualified glass edge grinding according to the type and purpose of the glass; Construct BP neural network to determine the maximum threshold of the variance of qualified glass after edge grinding; The analysis and judgment unit is used to comprehensively evaluate the surface roughness of the glass after the current edge grinding, and judge whether the edge grinding operation is needed; if the edge grinding operation is needed, control the glass edge grinding equipment to re-edge the glass that needs to be edged; If edge grinding is not required, the glass product is qualified and the edge grinding is completed.

Images