Multi-line laser combined device and microcrystalline glass detection method
Through the detection unit of the multi-line laser combined device, the problem of low detection efficiency of microcrystalline glass thickness, TTV and line trace is solved, and the accurate measurement of surface defects of microcrystalline glass is realized and the detection efficiency and accuracy are improved. sex.
Patent Information
- Application Number
- CN202411573290.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-06
- Publication Date
- 2024-12-06
AI Technical Summary
In the prior art, the thickness, TTV and line trace detection efficiency of microcrystalline glass is low, lacks accuracy, and the detection standards are subjective, making it difficult to achieve unified quality control standards.
Using a multi-line laser combined device, including a conveyor rack and multiple sets of detection units, each set of detection units consists of two first-line laser cameras and second-line laser cameras installed oppositely, and the surface of the microcrystalline glass is scanned by the translation adjustment unit to measure Line trace depth and thickness, and TTV is calculated by point cloud data alignment, thickness compensation and data fusion.
Quantitative detection of linear mark defects on the surface of microcrystalline glass and accurate measurement of thickness and TTV, reducing the impact of noise, improving detection efficiency and accuracy, and avoiding mutual interference between the laser lines of the main and secondary cameras.
Smart Images

Figure CN119086576A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer vision technology, and in particular to a multi-line laser combined device and a microcrystalline glass detection method. Background Art
[0002] Glass-ceramics uses a high-speed steel wire carrying abrasives attached to the steel wire to grind the glass block to achieve the purpose of cutting glass. In actual operation, the steel wire cannot be cut at the original position due to sudden shutdown and restart; the thickness and hardness of the welding point are different after the steel wire is disconnected; the adjustment of the load platform speed, spindle feed speed and speed parameters, etc., will inevitably cause various defects in the glass-ceramics. Among them, the three indicators of thickness, TTV (Total Thickness Variation) and line marks have a greater impact on the quality of glass-ceramics, and are currently an indispensable detection link for glass-ceramics manufacturers.
[0003] At present, there are few studies on the thickness, TTV and line mark detection of glass materials at home and abroad. The patent document "An optical glass surface roughness detection device" with publication number CN118067071A adopts a contact measurement method, and the roughness detection is performed by contacting the coating on the optical glass with a stylus. The patent document "A glass substrate appearance detection device" with publication number CN220913008U, the patent document "A method for detecting defects in curved glass" with publication number CN117990719A, and the patent document "Surface defect detection system based on microcrystalline glass panel finished products" with publication number CN117517348A all adopt traditional visual detection, that is, the image of the glass to be detected is obtained through an industrial camera, and the appearance defect detection is performed according to the grayscale of the pixel points. Since the industrial camera lacks height information, these methods cannot be used to measure the thickness, TTV and line marks of glass materials, and have no reference significance.
[0004] Some domestic manufacturers use manual inspection. The measuring tool for thickness and TTV is a handheld micrometer thickness gauge, which measures the thickness multiple times at different positions of the microcrystalline glass. Line mark detection does not use any measuring tools or optical observation instruments. It is detected by touching the glass surface with fingers to see if there is a undulating touch and observing with eyes to see if there are linear textures. The true depth of the line mark cannot be judged, and the detection efficiency is low. In addition, each person's judgment standard for the product is relatively subjective, making it difficult to reach a unified quality control standard, which can easily lead to inaccurate judgments. Summary of the invention
[0005] In view of the shortcomings of the prior art, the present invention provides a multi-line laser combined device and a microcrystalline glass detection method, which solves the problem of line mark detection without the aid of any measuring tools and optical observation instruments. It is impossible to judge whether there is an undulating touch on the glass surface by touching it with fingers and whether there are linear textures when observing with eyes. The actual depth of the line mark cannot be judged, and the detection efficiency is low. In addition, each person's judgment standard for the product is relatively subjective, and it is difficult to achieve a unified quality control standard, which can easily lead to inaccurate judgment.
[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a multi-line laser joint device, including a conveyor frame, a plurality of brackets and a plurality of detection units installed on the brackets are fixedly arranged above the conveyor frame, any group of detection units includes two first-line laser cameras and a second-line laser camera installed opposite to each other up and down, and a translation adjustment unit is arranged between the first-line laser camera and the second-line laser camera and the bracket;
[0007] Multiple groups of detection units are arranged separately along the conveying direction of the conveyor frame, and the multiple groups of detection units are staggered in the direction perpendicular to the conveyor frame direction. Multiple first-line laser cameras and multiple second-line laser cameras are respectively used to scan the contour data of the front and back sides of the microcrystalline glass, and can measure the true depth of the line marks. The first-line laser camera and the second-line laser camera are symmetrically installed and can measure the thickness and TTV of the microcrystalline glass after calibration.
[0008] Preferably, a method for detecting glass-ceramics using a multi-line laser combined device comprises the following steps:
[0009] Step 1: Align point cloud data of three groups of detection units;
[0010] Step 2: calibrate the thickness compensation values of three sets of detection units;
[0011] Step 3: Collect data;
[0012] Step 4: Calculate thickness, TTV, and line marks.
[0013] Preferably, the step of aligning the point cloud data by the three groups of detection units is specifically the following steps:
[0014] Step 1: The laser lines of the first laser camera and the second laser camera overlap;
[0015] Prepare a piece of glass with one smooth side and one rough side, with the rough side facing the first-line laser camera and the smooth side facing the second-line laser camera, and use the translation adjustment unit to adjust the vertical positions of the first-line laser camera and the second-line laser camera respectively to ensure that the glass is within the depth of field of the two cameras; monitor the contour line of the first-line laser camera, and two contour lines will appear in the display area, one is the laser line emitted by the first-line laser camera and received by the first-line laser camera, and the other is the laser line emitted by the second-line laser camera through the glass and received by the first-line laser camera, and use the translation adjustment unit to adjust the horizontal position of the second-line laser camera to move the second-line laser camera until the two contour lines overlap;
[0016] Step 2: Align the starting points of the first line laser camera and the second line laser camera;
[0017] Place the calibration block at the starting point of the laser line of the first-line laser camera so that the laser line of the first-line laser camera extends out of the calibration block. Then turn on the first-line laser camera and the second-line laser camera respectively to collect point cloud data. The first non-zero point is the index value of the starting point.
[0018] Step 3: Data valid interval;
[0019] Place the calibration block within the depth of field of the first-line laser camera and the second-line laser camera, so that the laser lines of the first-line laser camera and the second-line laser camera are completely on the calibration block, turn on the first-line laser camera and the second-line laser camera respectively, and collect point cloud data. The first valid point is the starting point of the valid data interval, and the last valid point is the end point of the valid data interval;
[0020] Step 4: Data trimming;
[0021] Set the main camera starting point index to , the secondary camera starting point index is , the valid range of main camera data is , The first valid point of the main camera point cloud data. The last valid point of the main camera point cloud data; the valid interval of the secondary camera data is , It is the first valid point of the secondary camera point cloud data. The last valid point of the secondary camera point cloud data; if the number of cropping points is , from the main camera Start cutting from a valid point;
[0022] The data range retained by the main camera after cropping is:
[0023] ,
[0024] The data range retained by the secondary camera is:
[0025] .
[0026] The multi-line laser joint device has three groups of detection units, and each group of detection units executes the above steps once to achieve point cloud data alignment.
[0027] Preferably, the step of calibrating the thickness compensation values of the three groups of detection units is specifically:
[0028] The thickness measurement formula after correcting the variables is:
[0029]
[0030] in:
[0031] H——thickness of the object being measured;
[0032] A——Measurement value of the main camera; (depth value, usually expressed as a larger integer)
[0033] B——Measurement value of the secondary camera; (depth value, usually expressed as a larger integer)
[0034] X——the coefficient of converting depth value to physical depth;
[0035] Y——thickness compensation value;
[0036] The coefficient X is automatically calibrated when the camera leaves the factory. This value does not need to be calibrated by the user. Therefore, in actual calibration, only one calibration plate is needed to calibrate the Y value. Assume that the thickness of the calibration plate is , the readings of the first-line laser camera and the second-line laser camera are , , then the thickness compensation value ; and there are three thickness compensation values for the three groups of detection units.
[0037] Preferably, the data collection step is specifically as follows:
[0038] The first-line laser camera and the second-line laser camera adopt a time-sharing triggering mode, that is, the first-line laser camera is triggered to collect data first, and the interval Only then is the second line laser camera triggered to collect data.
[0039] Preferably, the step of calculating thickness, TTV and line marks is specifically as follows:
[0040] There are 6 line laser cameras in total in the three detection units, and there are 6 depth maps, which are stored according to the first line laser camera depth map of the first detection unit, the second line laser camera depth map of the first detection unit, the first line laser camera depth map of the second detection unit, the second line laser camera depth map of the second detection unit, the first line laser camera depth map of the third detection unit, and the second line laser camera depth map of the third detection unit, and are recorded as , the specific steps are as follows:
[0041] Step 1: Each depth map in the data is filtered to obtain smooth data ;
[0042] Step 2: Each depth map in the calculation of the line traces in turn, a total of 6 line trace depth values are obtained , take the maximum value as the line mark depth of the current product;
[0043] Step 3: Each depth map in the data is fused in turn, and a total of 6 contour line data are obtained. ;
[0044] Step 4: and , and , and Generate thickness values respectively and get thickness ;
[0045] Step 5: The average value of all elements in is the thickness of the current product;
[0046] Step 6: Each set of thickness Divide into Then, we calculate the mean of the elements in each segment and get Mean , The difference between the maximum and minimum values in is the TTV of the current product.
[0047] Preferably, the data filtering method in step 1 is to filter the depth map Line by line Submedian filtering and Sub-mean filtering.
[0048] Preferably, the method for calculating line marks in step 2 has the following specific process:
[0049] Step 1: Depth Map The number of rows is , the line depth set of each row of data is , ;
[0050] Step 2: If , proceed to the next step, otherwise proceed to step 9;
[0051] Step 3: Get No. Row data, if the depth value is 0, it is considered as an invalid point and is removed;
[0052] Step 4: Remove the header and tail value, the The value is set to 100;
[0053] Step 5: Use the random sampling consistency algorithm to fit the straight line and perform horizontal correction;
[0054] Step 6: Set the window size to , starting from the first element, with a step size of Slide, and the difference between the maximum and minimum values of the elements in each window is used as the line depth of the current window, and several line depth values can be obtained;
[0055] Step 7: Sort the line depth values from largest to smallest. The average of the values is used as the line depth of the current row and added to the collection middle;
[0056] Step 8 , execute step 2;
[0057] Step 9: Arrange the The elements are sorted, and the average of the first two values is the depth map Depth of line marks.
[0058] Preferably, the data fusion method in step 3 has the following specific process:
[0059] Step 1: Depth Map The number of rows is , the number of columns is , two-dimensional array The elements are all 0, and the number of rows is , the number of columns is ,in is the number of points between two consecutive camera data collections. ;
[0060] Step 2: If , proceed to the next step, otherwise proceed to step 5;
[0061] Step 3: Get No. Row Data , and Assign to No. Row, No. Column to ;
[0062] Step 4 , execute step 2;
[0063] Step 5: Take the mean of the non-zero elements in each column, and get row 1 and column The array is the fused contour data .
[0064] Preferably, the method for generating the thickness value in step 4 is as follows:
[0065] Step 1: Set the main camera contour data to , the secondary camera contour data is , both are 1 line, An array of columns;
[0066] Step 2: If and If the elements in the same column are all non-zero, then both elements in the column will be retained, otherwise they will be removed to obtain new contour line data , the secondary camera contour data is , the number of rows and columns is the same;
[0067] Step 3: Remove the and Head and tail value, denoted as , , the number of rows and columns is the same;
[0068] Step 4: According to the formula Calculate the thickness.
[0069] The present invention discloses a multi-line laser combined device and a glass-ceramic detection method, which have the following beneficial effects:
[0070] The multi-line laser joint device and microcrystalline glass detection method use a joint device composed of 6 3D line laser cameras, which can detect and quantify microcrystalline glass surface line defects and measure thickness and TTV; microcrystalline glass is translucent, and the time-sharing trigger method can avoid mutual interference between the laser lines of the main and auxiliary cameras. The proposed data fusion method can suppress noise and reduce the impact of product jitter on thickness and TTV measurement. BRIEF DESCRIPTION OF THE DRAWINGS
[0071] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0072] Figure 1 It is a schematic diagram of the structure of the present invention;
[0073] Figure 2 It is a partial structural schematic diagram of the present invention;
[0074] Figure 3 It is a schematic diagram of the overall process of the present invention;
[0075] Figure 4 This is a thickness compensation principle diagram of the present invention;
[0076] Figure 5 Flow chart for calculating thickness, TTV and line marks of the present invention;
[0077] Figure 6 A flow chart of data filtering of the present invention;
[0078] Figure 7 A flow chart for calculating line marks of the present invention;
[0079] Figure 8 It is a flow chart of data fusion of the present invention;
[0080] Fig. 9 Flowchart for generating thickness values for the present invention.
[0081] In the figure: 1. conveyor frame; 2. bracket; 3. detection unit; 31. first line laser camera; 32. second line laser camera; 33. translation adjustment unit. DETAILED DESCRIPTION
[0082] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are 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.
[0083] The embodiment of the present application provides a multi-line laser joint device to solve the problem of line mark detection without the help of any measuring tools and optical observation instruments. It is impossible to judge the true depth of the line mark by touching the glass surface with fingers to see if there are undulating tactile sensations and observing with eyes to see if there are linear lines. The detection efficiency is low, and each person's judgment standard for the product is relatively subjective, making it difficult to achieve a unified quality control standard, which can easily lead to inaccurate judgments. The joint device composed of 6 3D line laser cameras can detect and quantify line mark defects on the surface of microcrystalline glass and measure thickness and TTV; microcrystalline glass is translucent, and the time-sharing triggering method can avoid mutual interference between the laser lines of the main and auxiliary cameras. The proposed data fusion method can suppress noise and reduce the impact of product jitter on thickness and TTV measurement.
[0084] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0085] The embodiments of the present invention disclose a multi-line laser combined device and a glass-ceramic detection method.
[0086] According to the attached Figure 1-2 As shown, a multi-line laser joint device includes a conveyor frame 1, and a plurality of brackets 2 and a plurality of detection units 3 installed on the brackets 2 are fixedly arranged above the conveyor frame 1, and any group of detection units 3 includes two first-line laser cameras 31 and second-line laser cameras 32 installed opposite to each other up and down, and a translation adjustment unit 33 is arranged between the first-line laser camera 31 and the second-line laser camera 32 and the bracket 2;
[0087] Multiple groups of detection units 3 are arranged separately along the conveying direction of the conveyor frame 1, and the multiple groups of detection units 3 are staggered in the direction perpendicular to the direction of the conveyor frame 1. Multiple first-line laser cameras 31 and multiple second-line laser cameras 32 are respectively used to scan the contour data of the front and back surfaces of the microcrystalline glass, and can measure the true depth of the line marks. The first-line laser camera 31 and the second-line laser camera 32 are symmetrically installed, and after calibration, they can measure the thickness and TTV of the microcrystalline glass.
[0088] According to the attached Figure 3 As shown, a method for detecting glass-ceramics using a multi-line laser combined device comprises the following steps:
[0089] Step 1: three groups of detection units 3 perform point cloud data alignment;
[0090] Step 2: calibrate the thickness compensation values of three groups of detection units 3;
[0091] Step 3: Collect data;
[0092] Step 4: Calculate thickness, TTV, and line marks.
[0093] The three groups of detection units 3 perform point cloud data alignment steps specifically as follows:
[0094] Step 1: The laser lines of the first line laser camera 31 and the second line laser camera 32 overlap;
[0095] Prepare a piece of glass with one side smooth and the other side rough, with the rough side facing the first line laser camera 31 and the smooth side facing the second line laser camera 32, and use the translation adjustment unit 33 to adjust the vertical positions of the first line laser camera 31 and the second line laser camera 32 respectively to ensure that the glass is within the depth of field of the two cameras; monitor the contour line of the first line laser camera 31, and two contour lines will appear in the display area, one is the laser line emitted by the first line laser camera 31 and received by the first line laser camera 31, and the other is the laser line emitted by the second line laser camera 32 through the glass and received by the first line laser camera 31, and use the translation adjustment unit 33 to adjust the horizontal position of the second line laser camera 32 to move the second line laser camera 32 until the two contour lines overlap;
[0096] Step 2: Align the starting points of the first line laser camera 31 and the second line laser camera 32;
[0097] Place the calibration block at the starting point of the laser line of the first line laser camera 31, so that the laser line of the first line laser camera 31 extends out of the calibration block, and then turn on the first line laser camera 31 and the second line laser camera 32 respectively to collect point cloud data. The first non-zero point is the index value of the starting point.
[0098] Step 3: Data valid interval;
[0099] Place the calibration block within the depth of field of the first line laser camera 31 and the second line laser camera 32, so that the laser lines of the first line laser camera 31 and the second line laser camera 32 are completely on the calibration block, turn on the first line laser camera 31 and the second line laser camera 32 respectively, and collect point cloud data. The first valid point is the starting point of the valid data interval, and the last valid point is the end point of the valid data interval;
[0100] Step 4: Data trimming;
[0101] Set the main camera starting point index to , the secondary camera starting point index is , the valid range of main camera data is , The first valid point of the main camera point cloud data. The last valid point of the main camera point cloud data; the valid interval of the secondary camera data is , It is the first valid point of the secondary camera point cloud data. The last valid point of the secondary camera point cloud data; if the number of cropping points is , from the main camera Start cutting from a valid point;
[0102] The data range retained by the main camera after cropping is:
[0103] ,
[0104] The data range retained by the secondary camera is:
[0105] .
[0106] The multi-line laser joint device has three groups of detection units 3, and each group of detection units 3 can realize point cloud data alignment by performing the above steps once.
[0107] According to the attached Figure 4 As shown, the steps of calibrating the thickness compensation values of the three groups of detection units 3 are as follows:
[0108] The thickness measurement formula after correcting the variables is:
[0109]
[0110] in:
[0111] H——thickness of the object being measured;
[0112] A——Measurement value of the main camera; (depth value, usually expressed as a larger integer)
[0113] B——Measurement value of the secondary camera; (depth value, usually expressed as a larger integer)
[0114] X——the coefficient of converting depth value to physical depth;
[0115] Y——thickness compensation value;
[0116] The coefficient X is automatically calibrated when the camera leaves the factory. This value does not need to be calibrated by the user. Therefore, in actual calibration, only one calibration plate is needed to calibrate the Y value. Assume that the thickness of the calibration plate is , the readings of the first line laser camera 31 and the second line laser camera 32 are respectively , , then the thickness compensation value ; and there are three thickness compensation values for the three groups of detection units 3.
[0117] The specific steps for collecting data are as follows:
[0118] The first line laser camera 31 and the second line laser camera 32 are triggered in a time-sharing manner, that is, the first line laser camera 31 is triggered to collect data first, and the interval Only then is the second line laser camera 32 triggered to collect data.
[0119] According to the attached Figure 5 As shown, the steps of calculating thickness, TTV and line marks are specifically as follows:
[0120] There are 6 line laser cameras in total in the three groups of detection units 3, and there are 6 depth maps, which are stored according to the depth map of the first line laser camera 31 of the first group of detection units 3, the depth map of the second line laser camera 32 of the first group of detection units 3, the depth map of the first line laser camera 31 of the second group of detection units 3, the depth map of the second line laser camera 32 of the second group of detection units 3, the depth map of the first line laser camera 31 of the third group of detection units 3, and the depth map of the second line laser camera 32 of the third group of detection units 3, and are recorded as , the specific steps are as follows:
[0121] Step 1: Each depth map in the data is filtered to obtain smooth data ;
[0122] Step 2: Each depth map in the calculation of the line traces in turn, a total of 6 line trace depth values are obtained , take the maximum value as the line mark depth of the current product;
[0123] Step 3: Each depth map in the data is fused in turn, and a total of 6 contour line data are obtained. ;
[0124] Step 4: and , and , and Generate thickness values respectively and get thickness ;
[0125] Step 5: The average value of all elements in is the thickness of the current product;
[0126] Step 6: Each set of thickness Divide into Then, we calculate the mean of the elements in each segment and get Mean , The difference between the maximum and minimum values in is the TTV of the current product.
[0127] According to the attached Figure 6 As shown, the data filtering method in step 1 is used to filter the depth map Line by line Submedian filtering and Sub-mean filtering.
[0128] According to the attached Figure 7 As shown, the method for calculating line marks in step 2 has the following specific process:
[0129] Step 1: Depth Map The number of rows is , the line depth set of each row of data is , ;
[0130] Step 2: If , proceed to the next step, otherwise proceed to step 9;
[0131] Step 3: Get No. Row data, if the depth value is 0, it is considered as an invalid point and is removed;
[0132] Step 4: Remove the header and tail value, the The value is set to 100;
[0133] Step 5: Use the random sampling consistency algorithm to fit the straight line and perform horizontal correction;
[0134] Step 6: Set the window size to , starting from the first element, with a step size of Slide, and the difference between the maximum and minimum values of the elements in each window is used as the line depth of the current window, and several line depth values can be obtained;
[0135] Step 7: Sort the line depth values from largest to smallest. The average of the values is used as the line depth of the current row and added to the collection middle;
[0136] Step 8 , execute step 2;
[0137] Step 9: Arrange the The elements are sorted, and the average of the first two values is the depth map Depth of line marks.
[0138] According to the attached Figure 8 As shown, the data fusion method in step 3 has the following specific process:
[0139] Step 1: Depth Map The number of rows is , the number of columns is , two-dimensional array The elements are all 0, and the number of rows is , the number of columns is ,in is the number of points between two consecutive camera data collections. ;
[0140] Step 2: If , proceed to the next step, otherwise proceed to step 5;
[0141] Step 3: Get No. Row Data , and Assign to No. Row, No. Column to ;
[0142] Step 4 , execute step 2;
[0143] Step 5: Take the mean of the non-zero elements in each column, and get row 1 and column The array is the fused contour data .
[0144] According to the attached Fig. 9 As shown, the method for generating thickness value in step 4 has the following specific process:
[0145] Step 1: Set the main camera contour data to , the secondary camera contour data is , both are 1 line, An array of columns;
[0146] Step 2: If and If the elements in the same column are all non-zero, then both elements in the column will be retained, otherwise they will be removed to obtain new contour line data , the secondary camera contour data is , the number of rows and columns is the same;
[0147] Step 3: Remove the and Head and tail value, denoted as , , the number of rows and columns is the same;
[0148] Step 4: According to the formula Calculate the thickness. Example
[0149] First, the point cloud data is aligned in four steps: the laser lines of the first laser camera 31 and the second laser camera 32 overlap, the starting points of the first laser camera 31 and the second laser camera 32 are aligned, the data valid interval, and the data is cropped. The data intervals retained by the six cameras are , , , , , .
[0150] Then, use a thickness of The mm calibration plate is used to calibrate the three groups of detection units 3. mm, provided by the line laser camera manufacturer, the thickness compensation value calibration result is: , , .
[0151] Afterwards, the first line laser camera 31 and the second line laser camera 32 collect data in a time-sharing triggering manner. Microseconds, based on The depth map is cropped and the first four images are horizontally flipped. Each image has 100 rows and 2000 columns.
[0152] Finally, the thickness, TTV, and line marks of the glass-ceramic are calculated. , the window size is 13, the number of mean filters , the window size is 31, and the 6 depth images are smoothed. Parameters related to line marks , window size , step length , ; Parameters of data fusion ; Parameters for generating thickness values ; TTV related parameters The calculated results of line mark, thickness and TTV are 0.0069mm, 0.6768mm and 0.0276mm respectively.
[0153] The combined device composed of 6 3D line laser cameras can detect and quantify surface line defects of microcrystalline glass and measure thickness and TTV; microcrystalline glass is translucent, and the time-sharing trigger method can avoid mutual interference between the laser lines of the main and auxiliary cameras. The proposed data fusion method can suppress noise and reduce the impact of product jitter on thickness and TTV measurement.
[0154] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. A multi-line laser combined device, comprising a conveyor frame (1), characterized in that: A plurality of brackets (2) and a plurality of detection units (3) mounted on the brackets (2) are fixedly arranged above the conveyor frame (1), each detection unit (3) comprising a first-line laser camera (31) and a second-line laser camera (32) mounted opposite to each other up and down, and a translation adjustment unit (33) is arranged between the first-line laser camera (31) and the second-line laser camera (32) and the bracket (2); The plurality of detection units (3) are arranged separately along the conveying direction of the conveying frame (1), and the plurality of detection units (3) are installed in a staggered manner in a direction perpendicular to the direction of the conveying frame (1). The plurality of first-line laser cameras (31) and the plurality of second-line laser cameras (32) are respectively used to scan the contour data of the front and back surfaces of the microcrystalline glass, and can measure the true depth of the line marks. The first-line laser cameras (31) and the second-line laser cameras (32) are installed opposite to each other, and can measure the thickness and TTV of the microcrystalline glass after calibration.
2. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 1, characterized in that: The following steps are involved: Step 1: aligning point cloud data of three groups of detection units (3); Step 2: calibrate the thickness compensation values of the three groups of detection units (3); Step 3: Collect data; Step 4: Calculate thickness, TTV, and line marks.
3. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 2, characterized in that: The three groups of detection units (3) perform point cloud data alignment steps specifically including the following steps: Step 1: The laser lines of the first laser camera (31) and the second laser camera (32) overlap; Prepare a piece of glass with one smooth side and one rough side, with the rough side facing the first line laser camera (31) and the smooth side facing the second line laser camera (32), and use the translation adjustment unit (33) to adjust the vertical positions of the first line laser camera (31) and the second line laser camera (32) respectively to ensure that the glass is within the depth of field of the two cameras; monitor the contour line of the first line laser camera (31), and two contour lines will appear in the display area, one is the laser line emitted by the first line laser camera (31) and received by the first line laser camera (31), and the other is the laser line emitted by the second line laser camera (32) passing through the glass and received by the first line laser camera (31), and use the translation adjustment unit (33) to adjust the horizontal position of the second line laser camera (32) to move the second line laser camera (32) until the two contour lines overlap; Step 2: aligning the starting points of the first line laser camera (31) and the second line laser camera (32); The calibration block is placed at the starting point of the laser line of the first line laser camera (31) so that the laser line of the first line laser camera (31) extends out of the calibration block, and then the first line laser camera (31) and the second line laser camera (32) are turned on respectively to collect point cloud data, and the first non-zero point is the index value of the starting point; Step 3: Data valid interval; Placing the calibration block within the depth of field of the first-line laser camera (31) and the second-line laser camera (32), so that the laser lines of the first-line laser camera (31) and the second-line laser camera (32) are completely on the calibration block, turning on the first-line laser camera (31) and the second-line laser camera (32) respectively, collecting point cloud data, the first valid point being the starting point of the valid data interval, and the last valid point being the end point of the valid data interval; Step 4: Data trimming; Set the main camera starting point index to , the secondary camera starting point index is , the valid range of main camera data is , The first valid point of the main camera point cloud data. The last valid point of the main camera point cloud data; the valid interval of the secondary camera data is , It is the first valid point of the secondary camera point cloud data. The last valid point of the secondary camera point cloud data; if the number of cropping points is , from the main camera Start cutting from a valid point; The data range retained by the main camera after cropping is: , The data range retained by the secondary camera is: 。 4. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 2, characterized in that: The steps of calibrating the thickness compensation values of the three groups of detection units (3) are as follows: The thickness measurement formula after correcting the variables is: , in: H——thickness of the object being measured; A – measurement of the main camera; B – measurement value of the secondary camera; X——the coefficient of converting depth value to physical depth; Y——thickness compensation value; The coefficient X is automatically calibrated when the camera leaves the factory. Assume that the thickness of the calibration plate is , the readings of the first line laser camera (31) and the second line laser camera (32) are , , then the thickness compensation value ; and there are three thickness compensation values for the three groups of detection units (3).
5. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 2, characterized in that: The data collection steps are specifically as follows: The first-line laser camera (31) and the second-line laser camera (32) are triggered in a time-sharing manner, that is, the first-line laser camera (31) is triggered to collect data, and the interval Only then is the second line laser camera (32) triggered to collect data.
6. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 2, characterized in that: The steps for calculating thickness, TTV and line marks are as follows: There are a total of 6 line laser cameras in the three detection units, and there are 6 depth maps, which are stored according to the depth map of the first line laser camera (31) of the first detection unit (3), the depth map of the second line laser camera (32) of the first detection unit (3), the depth map of the first line laser camera (31) of the second detection unit (3), the depth map of the second line laser camera (32) of the second detection unit (3), the depth map of the first line laser camera (31) of the third detection unit (3), and the depth map of the second line laser camera (32) of the third detection unit (3), and are recorded as , the specific steps are as follows: Step 1: Each depth map in the data is filtered to obtain smooth data ; Step 2: Each depth map in the calculation of the line traces in turn, a total of 6 line trace depth values are obtained , take the maximum value as the line mark depth of the current product; Step 3: Each depth map in the data is fused in turn, and a total of 6 contour line data are obtained. ; Step 4: and , and , and Generate thickness values respectively and get thickness ; Step 5: The average value of all elements in is the thickness of the current product; Step 6: Each set of thickness Divide into Then, we calculate the mean of the elements in each segment and get Mean , The difference between the maximum and minimum values in is the TTV of the current product.
7. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 6, characterized in that: The data filtering method in step 1 is used to filter the depth map Line by line Submedian filtering and Sub-mean filtering.
8. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 6, characterized in that: The specific process of the line mark calculation method in step 2 is as follows: Step 1: Depth Map The number of rows is , the line depth set of each row of data is , ; Step 2: If , proceed to the next step, otherwise proceed to step 9; Step 3: Get No. Row data, if the depth value is 0, it is considered as an invalid point and is removed; Step 4: Remove the header and tail value, the The value is set to 100; Step 5: Use the random sampling consistency algorithm to fit the straight line and perform horizontal correction; Step 6: Set the window size to , starting from the first element, with a step size of Slide, and the difference between the maximum and minimum values of the elements in each window is used as the line depth of the current window, and several line depth values can be obtained; Step 7: Sort the line depth values from largest to smallest. The average of the values is used as the line depth of the current row and added to the collection middle; Step 8 , execute step 2; Step 9: Arrange the The elements are sorted, and the average of the first two values is the depth map Depth of line marks.
9. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 6, characterized in that: The data fusion method in step 3 has the following specific process: Step 1: Depth Map The number of rows is , the number of columns is , two-dimensional array The elements are all 0, and the number of rows is , the number of columns is ,in is the number of points between two consecutive camera data collections. ; Step 2: If , proceed to the next step, otherwise proceed to step 5; Step 3: Get No. Row Data , and Assign to No. Row, No. Column to ; Step 4 , execute step 2; Step 5: Take the mean of the non-zero elements in each column, and get row 1 and column The array is the fused contour data .
10. The method for detecting glass-ceramics using a multi-line laser combined device according to claim 6, characterized in that: The method for generating thickness value in step 4 is as follows: Step 1: Set the main camera contour data to , the secondary camera contour data is , both are 1 line, An array of columns; Step 2: If and If the elements in the same column are all non-zero, then both elements in the column will be retained, otherwise they will be removed to obtain new contour line data , the secondary camera contour data is , the number of rows and columns is the same; Step 3: Remove the and Head and tail value, denoted as , , the number of rows and columns is the same; Step 4: According to the formula Calculate the thickness.
Citation Information
Patent Citations
Plate thickness detection method and device
CN115406367A
Line laser thickness measuring method, device and equipment and storage medium
CN118274726A
Calibration method of measuring device, measuring method, measuring device, medium and program product
CN118376166A
Thickness measuring system
KR101413511B1
Detection System Based on Modulation of Line Structured Laser Image of Glass
US20150226682A1
