A method for measuring the dimensions of a photovoltaic glass
Patent Information
- Application Number
- CN202611287871.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-24
- Publication Date
- 2026-10-23
AI Technical Summary
[0005]第二、传统固定像素当量、固定角度补偿或整图统一坐标变换通常将整幅扫描图像看作一个整体进行补偿，难以描述玻璃在扫描过程中的行级动态运动状态
[0053]本发明相对于现有技术的有益效果是：本发明利用同一扫描行左右边缘点构建双边缘中心轨迹，并由中心轨迹偏差反推行级横向偏摆量，能够更准确描述大尺寸光伏玻璃在线输送过程中的动态偏摆；另外，通过建立包含横向偏摆量、纵向残差和局部旋转角的行级刚体运动状态，对每一扫描行边缘点进行动态坐标补偿，能够降低玻璃输送偏摆、局部旋转、辊道打滑和编码器残差对线阵扫描图像造成的几何畸变；然后，利用玻璃宽度恒定、对边平行、邻边垂直、中心轨迹平滑和补偿参数平滑等约束进行联合优化，能够抑制边缘反光、局部毛刺和随机噪声导致的边缘点异常，提高长边、短边和对角线差测量的重复稳定性；最后充分利用双线阵相机的左右边缘同步测量信息，不依赖智能预测模型作为必要条件，也不以完整图像拼接作为必要前提，适合大尺寸光伏玻璃在线连续测量场景。
Smart Images
Figure CN122961298A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of dimensional measurement, and in particular to a method for measuring the dimensions of photovoltaic glass. Background Technology
[0002] Large-size photovoltaic glass is a key basic material in photovoltaic module manufacturing. Its geometric dimensional parameters, such as long side, short side, edge angle, and diagonal difference, directly affect the quality of subsequent coating, lamination, encapsulation, and module assembly. As photovoltaic glass develops towards large-format, high-speed, and online continuous production, traditional manual sampling or offline measurement methods are no longer sufficient to meet the real-time, stability, and consistency requirements of high-speed production lines.
[0003] Existing online methods for measuring the dimensions of large-size glass typically calculate glass dimensions through pixel equivalent conversion, encoder line spacing conversion, camera calibration, image stitching, edge detection, line fitting, and geometric constraints. While these methods can perform basic dimensional measurements, they still present the following challenges in high-speed transport and large-size sheet applications:
[0004] First, slight lateral sway, local rotation, roller slippage, encoder residuals, and trigger delays may occur during the glass conveyor process. Since linear array images are formed by scanning line by line, the motion error of the glass will accumulate along the scanning line direction, causing slight changes in the lateral position and local angle of the same piece of glass on different lines.
[0005] Secondly, traditional methods such as fixed pixel equivalent, fixed angle compensation, or unified coordinate transformation of the entire image typically treat the entire scanned image as a whole for compensation, making it difficult to describe the row-level dynamic motion of the glass during the scanning process. When the long side of the glass reaches the meter level, small angular deviations or minute lateral sway will be amplified by the long side dimension, causing fluctuations in the measurement results of the long side, short side, and diagonal difference.
[0006] Third, ordinary straight line fitting and parallel and perpendicular geometric constraints mainly perform post-processing on the edges at the measurement result level. They lack modeling of the sources of motion error during linear array scanning and are easily affected by local edge reflection, grayscale transition, edge burrs and transport jitter.
[0007] Therefore, there is a need for a dimensional measurement method that can utilize the synchronous information from the left and right edges of a dual-line array camera to reflect row-level sway and local rotation during the online transport of glass, and to dynamically compensate for each scan row, in order to improve the accuracy and repeatability of online dimensional measurement of large-size photovoltaic glass. Summary of the Invention
[0008] To address the aforementioned problems, this invention provides a dimension measurement method with high accuracy and strong stability.
[0009] To achieve the above objectives, the present invention provides a method for measuring the dimensions of photovoltaic glass, comprising: acquiring linear array scan images of the left and right regions of the photovoltaic glass to be measured during the transport process using a first linear array camera and a second linear array camera arranged in sections along the width direction of the photovoltaic glass to be measured, and simultaneously acquiring encoder count values corresponding to the scan lines; extracting glass edge points from the linear array scan images acquired by the first and second linear array cameras, and establishing the initial physical coordinates of the glass edge points based on the lateral pixel equivalent and the encoder count values; mapping the left and right edge points near the same scan line or the same encoder position, calculating the double edge center position of each scan line, and constructing a double edge center trajectory; and calculating the double edge center trajectory based on the double edge center trajectory and the ideal center trajectory. The deviation between tracks is used to estimate the lateral yaw amount corresponding to each scan line; the local rotation angle corresponding to each scan line is estimated based on the local slope of the left and right edges; and the longitudinal residual is estimated based on encoder position feedback and glass geometric constraints, thereby establishing a row-level rigid body motion state that includes lateral yaw amount, longitudinal residual, and local rotation angle; based on the row-level rigid body motion state, row-level rigid body motion compensation is performed on the initial physical coordinates of the edge points of the corresponding scan line to obtain the compensated edge point coordinates; the compensated edge point coordinates satisfy the constraint conditions, which include one or more of the following: edge straight line consistency, constant width, center trajectory smoothing, and compensation parameter smoothing; based on the compensated edge point coordinates, the actual length, actual width, edge angle, and diagonal difference of the photovoltaic glass under test are calculated and output.
[0010] Optionally, the initial physical coordinates include initial horizontal physical coordinates and initial vertical physical coordinates. The initial horizontal physical coordinates are obtained by converting the horizontal pixel coordinates of the first / second linear array camera, and the initial vertical physical coordinates are obtained by converting the encoder count value or the scan line number.
[0011] Optionally, the dual-edge center trajectory is calculated using the following formula:
[0012] ;
[0013] in, This indicates the horizontal position of the center of the double edges in the nth row; This represents the horizontal physical coordinates of the left edge point of the nth row; This represents the horizontal physical coordinates of the right edge point of the nth row.
[0014] Optionally, the ideal center trajectory is represented by the following formula:
[0015] ;
[0016] in, denoted by , a represents the lateral position of the ideal center trajectory corresponding to the nth row; denoted by , a represents the slope parameter of the ideal center trajectory relative to the longitudinal physical coordinates; denoted by , b represents the lateral intercept parameter of the ideal center trajectory. Represents the initial vertical physical coordinates of the nth row;
[0017] The lateral yaw rate is obtained by the following formula:
[0018] ;
[0019] in, This represents the lateral yaw rate of the nth row; This indicates the horizontal position of the ideal center trajectory in the nth row.
[0020] Optionally, the local rotation angle is obtained by jointly estimating the local slopes of the left and right edges, specifically as follows:
[0021] ;
[0022] in, This represents the local rotation angle corresponding to the nth row: This represents the local slope of the left edge near the nth row; This represents the local slope of the right edge near the nth row; Represents the arctangent function;
[0023] The local slopes of the left and right edges are obtained as follows:
[0024] ;
[0025] ;
[0026] in, , y represents the horizontal physical coordinates of the left and right edges, respectively; y represents the vertical physical coordinate.
[0027] Optionally, the row-level rigid body motion state is represented by the following vector:
[0028] ;
[0029] in, Represents the row-level rigid body motion state vector of the nth row; This represents the local rotation angle of the nth row; This represents the vertical residual of the nth row;
[0030] The longitudinal residual satisfy:
[0031] ;
[0032] Where y(n) represents the nth row of vertical physical coordinates after vertical residual correction; This represents the actual conveying displacement corresponding to a unit encoder count; C(n) represents the encoder count value corresponding to the nth row of data acquisition. Indicates the reference encoder count value;
[0033] Optionally, based on the row-level rigid body motion state, row-level rigid body motion compensation is performed on the initial physical coordinates of the edge points of the corresponding scan row to obtain the compensated edge point coordinates, including: specifically obtaining the compensated coordinates through the following formula:
[0034] ;
[0035] in, and These represent the compensated horizontal and vertical physical coordinates, respectively. Indicates the initial horizontal physical coordinates before compensation; Indicates the local rotation angle Rotation compensation matrix in opposite directions;
[0036] The rotation compensation matrix is:
[0037] .
[0038] Optionally, the edge line consistency constraint is:
[0039] ;
[0040] in, Indicates the error in the consistency of edge straight lines; This represents the j-th compensated edge point; This represents the fitted straight line corresponding to the glass edge to which the j-th edge point belongs; , ) represents a point To the fitted line The vertical distance;
[0041] The constant width constraint is:
[0042] ;
[0043] in, This indicates a constant width error; This represents the horizontal physical coordinates of the right edge after compensation in the nth row; Represents the horizontal physical coordinates of the left edge after compensation in the nth row; W represents the actual width or estimated width of the photovoltaic glass to be measured.
[0044] The center trajectory smoothing constraint is:
[0045] ;
[0046] in, This indicates the smoothing error of the center trajectory; This represents the second-order change of the double-edge center trajectory with respect to the scan row number;
[0047] The compensation parameter smoothing constraint is:
[0048] ;
[0049] in, This indicates the smoothing error of the compensation parameters; This represents the second derivative of the lateral yaw with respect to the scan row number; This represents the second derivative of the longitudinal residual with respect to the scan row number; This represents the second derivative of the local rotation angle with respect to the scan row number; , and These represent the weighting coefficients for the lateral sway, longitudinal residual, and local rotation angle smoothing term, respectively.
[0050] Optionally, the diagonal difference is obtained using the following formula:
[0051] ;
[0052] in, Indicates the difference between the diagonals; This represents the length of the first diagonal calculated based on the first pair of diagonal points after compensation. This represents the length of the second diagonal calculated based on the compensated second pair of diagonal points.
[0053] The advantages of this invention compared to existing technologies are as follows: This invention utilizes the left and right edge points of the same scanning row to construct a dual-edge center trajectory, and uses the center trajectory deviation to inversely calculate the lateral sway of the linear array, which can more accurately describe the dynamic sway during the online transport of large-size photovoltaic glass. Furthermore, by establishing a row-level rigid body motion state including lateral sway, longitudinal residual, and local rotation angle, dynamic coordinate compensation is performed on the edge points of each scanning row, reducing the geometric distortion caused by glass transport sway, local rotation, roller slippage, and encoder residuals on the linear array scanning image. Then, by using constraints such as constant glass width, parallel opposite sides, perpendicular adjacent sides, smooth center trajectory, and smooth compensation parameters for joint optimization, edge point anomalies caused by edge reflection, local burrs, and random noise can be suppressed, improving the repeatability stability of long side, short side, and diagonal difference measurements. Finally, it fully utilizes the synchronous measurement information of the left and right edges of the dual linear array camera, without relying on intelligent prediction models as a necessary condition or requiring complete image stitching as a prerequisite, making it suitable for online continuous measurement scenarios of large-size photovoltaic glass. Attached Figure Description
[0054] Figure 1 This is a flowchart of a method for measuring the dimensions of photovoltaic glass provided by the present invention;
[0055] Figure 2 This is a schematic diagram of the dual-line array camera and the dual-edge center trajectory provided by the present invention;
[0056] Figure 3 This is a schematic diagram of the row-level rigid body motion compensation model provided by the present invention;
[0057] Figure 4 This is a schematic diagram of the joint optimization solution constraints provided by the present invention. Detailed Implementation
[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0059] Reference Figure 1 This embodiment provides a method for measuring the dimensions of photovoltaic glass, including the following steps:
[0060] S10: Acquire linear scan images of the photovoltaic glass to be inspected, and simultaneously obtain encoder count values corresponding to the scan lines.
[0061] Specifically, by using a first linear array camera and a second linear array camera arranged in sections along the width of the photovoltaic glass to be tested, linear array scan images of the left and right regions of the photovoltaic glass to be tested are acquired respectively during the transport of the photovoltaic glass, and encoder count values corresponding to the scan lines are acquired simultaneously.
[0062] S20: Establish the initial physical coordinates of the edge points.
[0063] Specifically, glass edge points are extracted from the linear scan images acquired by the first and second linear scan cameras, respectively. For the edge point obtained by the i-th linear scan camera in the n-th row, initial horizontal physical coordinates are first established based on the horizontal pixel equivalent:
[0064] ;
[0065] in, This represents the initial horizontal physical coordinates of the i-th line array camera in the n-th row; i=1,2, representing the first and second line array cameras respectively. This represents the lateral pixel coordinates of the edge point extracted by the i-th linear scan camera in the n-th row; Represents the lateral reference pixel coordinates of the i-th linear array camera; This represents the lateral pixel equivalent coefficient of the i-th linear array camera; This represents the lateral offset of the i-th linear array camera relative to the unified physical measurement coordinate system.
[0066] Establish the initial longitudinal physical coordinates based on the encoder count value:
[0067] ;
[0068] in, This represents the initial vertical physical coordinate corresponding to the nth row; This represents the actual conveying displacement corresponding to a unit encoder count. This represents the encoder count value corresponding to the nth row of data acquisition. This indicates the reference encoder count value.
[0069] S30: Construct the dual-edge center trajectory.
[0070] Specifically, refer to Figure 2 For left and right edge points near the same scan line or the same encoder position, calculate the center position of their double edges. All scanned lines Forming a double-edge center trajectory.
[0071] By mapping the left and right edge points near the same scan line or encoder position, the horizontal physical coordinates of the left edge can be obtained. and the horizontal physical coordinates of the right edge And calculate the center position of the double edges in the nth row:
[0072] ;
[0073] in, This indicates the horizontal position of the center of the double edges in the nth row; This represents the horizontal physical coordinates of the left edge point of the nth row; This represents the horizontal physical coordinates of the right edge point of the nth row.
[0074] Through all scan lines This forms a double-edge center trajectory that varies along the scanning direction. This center trajectory is used to characterize the lateral movement of the glass under test relative to the ideal transport direction during transport.
[0075] S40: Estimate the row-level lateral yaw, local rotation angle, and longitudinal residual to establish the row-level rigid body motion state, which includes the lateral yaw, longitudinal residual, and local rotation angle.
[0076] Continue to refer to Figure 2 If the glass is transported smoothly along the ideal direction, the center trajectory should be close to a smooth straight line; if the glass experiences lateral sway during transport, the center trajectory will deviate from the ideal center trajectory. This deviation is used as the lateral sway amount for the nth row. .
[0077] Specifically, the ideal center trajectory is obtained by performing linear fitting or low-order smooth fitting on the double-edge center trajectory:
[0078] ;
[0079] in, denoted by , a represents the lateral position of the ideal center trajectory corresponding to the nth row; denoted by , a represents the slope parameter of the ideal center trajectory relative to the longitudinal physical coordinates; denoted by , b represents the lateral intercept parameter of the ideal center trajectory. This represents the initial vertical physical coordinates of the nth row.
[0080] Subtracting the actual double-edge center trajectory from the ideal center trajectory yields the lateral sway of the nth row: ;
[0081] in, This represents the lateral yaw rate of the nth row; This indicates the horizontal position of the ideal center trajectory in the nth row.
[0082] The local rotation angle of the nth row is estimated based on the local slopes of the left and right edges. The local slopes of the left and right edges are as follows:
[0083] ;
[0084] ;
[0085] in, , y represents the horizontal physical coordinates of the left and right edges, respectively; y represents the vertical physical coordinate. This represents the local slope of the right edge near the nth row; This represents the local slope of the left edge near the nth row.
[0086] The local rotation angle of the nth row is:
[0087] ;
[0088] in, This represents the local rotation angle corresponding to the nth row; Represents the arctangent function;
[0089] Additionally, refer to Figure 3 Based on the obtained lateral sway, this embodiment further estimates the local rotation angle of the nth row according to the local slope of the left and right edges. And introduce longitudinal residuals This describes the minute difference between the encoder count and the actual longitudinal displacement of the glass. This constitutes the row-level rigid body motion state vector of the nth row. .
[0090] Specifically, introduce longitudinal residuals Describes the minute discrepancies between encoder counting, roller conveyor, and actual glass displacement:
[0091] ;
[0092] Where y(n) represents the nth row of vertical physical coordinates after vertical residual correction; This represents the actual conveying displacement corresponding to a unit encoder count; C(n) represents the encoder count value corresponding to the nth row of data acquisition. Indicates the reference encoder count value; This represents the vertical residual of the nth row.
[0093] Therefore, the row-level rigid body motion state of the nth row can be represented as:
[0094] ;
[0095] in, Represents the row-level rigid body motion state vector of the nth row; This represents the local rotation angle of the nth row.
[0096] S50: Based on the row-level rigid body motion state, perform row-level rigid body motion compensation on the initial physical coordinates of the edge points of the corresponding scan row to obtain the compensated edge point coordinates.
[0097] Specifically, based on the rigid body motion state of the nth row, the initial physical coordinates of the edge points of that row are dynamically compensated:
[0098] ;
[0099] in, and These represent the compensated horizontal and vertical physical coordinates, respectively. Indicates the initial horizontal physical coordinates before compensation; Indicates the local rotation angle Rotation compensation matrix in opposite directions;
[0100] The rotation compensation matrix is:
[0101] .
[0102] Furthermore, since the glass conveying sway and local rotation generally change continuously and slowly, the lateral sway, longitudinal residual, and local rotation angle can be expressed as smooth functions. In this embodiment, B-spline basis functions are used:
[0103] ;
[0104] ;
[0105] ;
[0106] in, This represents the function value of the m-th B-spline basis function at the n-th row; , and represents the B-spline coefficients corresponding to the lateral yaw, longitudinal residual, and local rotation angle, respectively; M represents the number of B-spline basis functions.
[0107] It should be noted that, in this embodiment, the compensated edge point coordinates satisfy the constraint conditions, which include one or more of the following: edge line consistency, constant width, smooth center trajectory, and smooth compensation parameters.
[0108] The edge line consistency constraint is as follows:
[0109] ;
[0110] in, Indicates the error in the consistency of edge straight lines; This represents the j-th compensated edge point; This represents the fitted straight line corresponding to the glass edge to which the j-th edge point belongs; , ) represents a point To the fitted line The vertical distance;
[0111] The constant width constraint is:
[0112] ;
[0113] in, This indicates a constant width error; This represents the horizontal physical coordinates of the right edge after compensation in the nth row; Represents the horizontal physical coordinates of the left edge after compensation in the nth row; W represents the actual width or estimated width of the photovoltaic glass to be measured.
[0114] The center trajectory smoothing constraint is:
[0115] ;
[0116] in, This indicates the smoothing error of the center trajectory; This represents the second-order change of the double-edge center trajectory with respect to the scan row number;
[0117] The compensation parameter smoothing constraint is:
[0118] ;
[0119] in, This indicates the smoothing error of the compensation parameters; This represents the second derivative of the lateral yaw with respect to the scan row number; This represents the second derivative of the longitudinal residual with respect to the scan row number; This represents the second derivative of the local rotation angle with respect to the scan row number; , and These represent the actual lengths of the weighting coefficients for the lateral sway, longitudinal residual, and local rotation angle smoothing term, respectively.
[0120] Reference Figure 4Finally, a joint optimization objective function is established:
[0121] ;
[0122] in, Indicates the error in the consistency of edge straight lines; This indicates a constant width error; This indicates the smoothing error of the center trajectory; This indicates the smoothing error of the compensation parameters; , and represents the weight coefficients of the constant width constraint, the center trajectory smoothing constraint, and the compensation parameter smoothing constraint in the joint optimization objective, respectively.
[0123] S60: Calculate and output the actual length, actual width, edge angle, and diagonal difference of the photovoltaic glass under test based on the compensated edge point coordinates.
[0124] Specifically, after completing the row-level rigid body motion compensation, the edge points are classified to obtain the left edge point set, right edge point set, front edge point set, and rear edge point set of the photovoltaic glass under test. Linear fitting is then performed on these four sets of edge points to obtain the left edge line, right edge line, front edge line, and rear edge line. By finding the intersection of adjacent edge lines, the four compensated physical corner points of the photovoltaic glass under test are obtained, denoted as: , , , ,in, , , , The four compensated physical corner points of the photovoltaic glass under test are represented sequentially in a clockwise or counterclockwise direction. In one embodiment, The left front corner point, The right front corner point, The right rear corner point, It is the left rear corner point.
[0125] The distance between any two compensated corner points is expressed as:
[0126] ;
[0127] The actual length and width of the photovoltaic glass under test are calculated based on the distances between the four compensated corner points. Actual length This can be obtained by averaging the lengths of the two longer sides:
[0128] ;
[0129] Actual width It can be obtained by averaging the lengths of the two shorter sides:
[0130] ;
[0131] in, This indicates the actual length of the photovoltaic glass being tested. This indicates the actual width of the photovoltaic glass to be tested.
[0132] The glass edge angle can be calculated based on the compensated long side direction. Taking the tilt angle of the long side relative to the glass conveying direction as an example, the angles of the left and right long sides are expressed as follows:
[0133] ;
[0134] ;
[0135] The overall edge angle of the photovoltaic glass under test can then be expressed as:
[0136] ;
[0137] in, This indicates the tilt angle of the left long side relative to the glass conveying direction. This indicates the tilt angle of the right long side relative to the glass conveying direction. This indicates the overall edge angle of the photovoltaic glass under test.
[0138] The lengths of the two diagonals are expressed as follows:
[0139] ;
[0140] ;
[0141] The difference between the diagonals can be expressed as:
[0142] ;
[0143] in, Indicates the difference between the diagonals; This represents the length of the first diagonal calculated based on the first pair of diagonal points after compensation. This represents the length of the second diagonal calculated based on the compensated second pair of diagonal points. The diagonal difference is used to evaluate the rectangularity and shape accuracy of the glass and cannot be replaced solely by theoretically calculated length and width values.
[0144] This invention utilizes the left and right edge points of the same scanning row to construct a dual-edge center trajectory, and uses the deviation of the center trajectory to inversely calculate the lateral sway of the linear array, which can more accurately describe the dynamic sway during the online transport of large-size photovoltaic glass. Furthermore, by establishing a row-level rigid body motion state including lateral sway, longitudinal residual, and local rotation angle, dynamic coordinate compensation is performed on the edge points of each scanning row, reducing the geometric distortion caused by glass transport sway, local rotation, roller slippage, and encoder residuals on the linear array scanning image. Then, joint optimization using constraints such as constant glass width, parallel opposite sides, perpendicular adjacent sides, smooth center trajectory, and smooth compensation parameters can suppress edge point anomalies caused by edge reflection, local burrs, and random noise, improving the repeatability stability of long side, short side, and diagonal difference measurements. Finally, it fully utilizes the synchronous measurement information of the left and right edges of the dual linear array camera, without relying on intelligent prediction models as a necessary condition or requiring complete image stitching as a prerequisite, making it suitable for online continuous measurement scenarios of large-size photovoltaic glass.
[0145] In addition, embodiments of the present invention also provide a computer-readable storage medium, wherein the computer-readable storage medium may store a program that, when executed, includes some or all of the steps of any photovoltaic glass size measurement method described in the above method embodiments.
[0146] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0147] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.
[0148] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.
[0149] An exemplary flowchart for implementing a method for measuring the dimensions of photovoltaic glass according to an embodiment of the present invention has been described above with reference to the accompanying drawings. It should be noted that the numerous details included in the above description are merely illustrative of the invention and not intended to limit it. In other embodiments of the invention, the method may have more, fewer, or different steps, and the order, inclusion, function, and other relationships between the steps may differ from those described and illustrated.
Claims
1. A method for measuring the dimensions of photovoltaic glass, characterized in that, include: By using a first linear array camera and a second linear array camera arranged in sections along the width of the photovoltaic glass to be tested, linear array scan images of the left and right regions of the photovoltaic glass to be tested are acquired respectively during the transport of the photovoltaic glass, and encoder count values corresponding to the scan lines are acquired simultaneously. Glass edge points are extracted from the linear scan images acquired by the first and second linear scan cameras, respectively, and the initial physical coordinates of the glass edge points are established based on the lateral pixel equivalent and encoder count value. By mapping the left and right edge points near the same scan line or the same encoder position, the double edge center position of each scan line is calculated, and the double edge center trajectory is constructed. Based on the deviation between the double-edge center trajectory and the ideal center trajectory, the lateral yaw amount corresponding to each scan line is estimated; the local rotation angle corresponding to each scan line is estimated based on the local slope of the left and right edges; and the longitudinal residual is estimated based on the encoder position feedback and glass geometric constraints, thereby establishing a row-level rigid body motion state that includes lateral yaw amount, longitudinal residual and local rotation angle. Based on the row-level rigid body motion state, row-level rigid body motion compensation is performed on the initial physical coordinates of the edge points of the corresponding scan row to obtain the compensated edge point coordinates; the compensated edge point coordinates satisfy the constraint conditions, which include one or more of the following: edge line consistency, constant width, center trajectory smoothing, and compensation parameter smoothing. The actual length, actual width, edge angle, and diagonal difference of the photovoltaic glass under test are calculated and output based on the compensated edge point coordinates.
2. The size measurement method according to claim 1, characterized in that, The initial physical coordinates include initial horizontal physical coordinates and initial vertical physical coordinates. The initial horizontal physical coordinates are obtained by converting the horizontal pixel coordinates of the first / second linear array camera, and the initial vertical physical coordinates are obtained by converting the encoder count value or the scan line number.
3. The size measurement method according to claim 1, characterized in that, The trajectory of the double edge center is calculated using the following formula: ； in, This indicates the horizontal position of the center of the double edges in the nth row; This represents the horizontal physical coordinates of the left edge point of the nth row; This represents the horizontal physical coordinates of the right edge point of the nth row.
4. The size measurement method according to claim 1, characterized in that, The ideal center trajectory is represented by the following formula: ； in, denoted by , a represents the lateral position of the ideal center trajectory corresponding to the nth row; denoted by , a represents the slope parameter of the ideal center trajectory relative to the longitudinal physical coordinates; denoted by , b represents the lateral intercept parameter of the ideal center trajectory. Represents the initial vertical physical coordinates of the nth row; The lateral yaw rate is obtained by the following formula: ； in, This represents the lateral yaw rate of the nth row; This indicates the horizontal position of the ideal center trajectory in the nth row.
5. The size measurement method according to claim 1, characterized in that, The local rotation angle is obtained by jointly estimating the local slopes of the left and right edges, specifically as follows: ； in, This represents the local rotation angle corresponding to the nth row: This represents the local slope of the left edge near the nth row; This represents the local slope of the right edge near the nth row; Represents the arctangent function; The local slopes of the left and right edges are obtained in the following ways: ； ； in, , y represents the horizontal physical coordinates of the left and right edges, respectively; y represents the vertical physical coordinate.
6. The dimension measurement method according to claim 1, characterized in that, The row-level rigid body motion state is represented by the following vector: ； in, Represents the row-level rigid body motion state vector of the nth row; This represents the local rotation angle of the nth row; This represents the vertical residual of the nth row; The longitudinal residual satisfy: ； Where y(n) represents the vertical physical coordinate of the nth row after correction of the vertical residual; This represents the actual conveying displacement corresponding to a unit encoder count; C(n) represents the encoder count value corresponding to the nth row of data acquisition. This indicates the reference encoder count value.
7. The dimension measurement method according to claim 6, characterized in that, Based on the row-level rigid body motion state, row-level rigid body motion compensation is performed on the initial physical coordinates of the edge points of the corresponding scan row to obtain the compensated edge point coordinates, including: specifically, the compensated coordinates are obtained through the following formula: ； in, and These represent the compensated horizontal and vertical physical coordinates, respectively. Indicates the initial horizontal physical coordinates before compensation; Indicates the local rotation angle Rotation compensation matrix in opposite directions; The rotation compensation matrix is: 。 8. The dimension measurement method according to claim 1, characterized in that, The edge line consistency constraint is: ； in, Indicates the error in the consistency of edge straight lines; This represents the j-th compensated edge point; This represents the fitted straight line corresponding to the glass edge to which the j-th edge point belongs; , ) represents a point To the fitted line The vertical distance; The constant width constraint is: ； in, Indicates a constant width error; This represents the horizontal physical coordinates of the right edge after compensation in the nth row; Represents the horizontal physical coordinates of the left edge after compensation in the nth row; W represents the actual width or estimated width of the photovoltaic glass to be measured. The center trajectory smoothing constraint is: ； in, This indicates the smoothing error of the center trajectory; This represents the second-order change of the double-edge center trajectory with respect to the scan row number; The compensation parameter smoothing constraint is: ； in, This indicates the smoothing error of the compensation parameters; This represents the second derivative of the lateral yaw with respect to the scan row number; This represents the second derivative of the longitudinal residual with respect to the scan row number; This represents the second derivative of the local rotation angle with respect to the scan row number; , and These represent the weighting coefficients for the lateral sway, longitudinal residual, and local rotation angle smoothing term, respectively.
9. The dimension measurement method according to claim 1, characterized in that, The difference between the diagonals is obtained using the following formula: ； in, Indicates the difference between the diagonals; This represents the length of the first diagonal calculated based on the first pair of diagonal points after compensation. This represents the length of the second diagonal calculated based on the compensated second pair of diagonal points.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of a method for measuring the dimensions of photovoltaic glass as described in any one of claims 1 to 9.