Photovoltaic cell piece edge grabbing and positioning processing method and system

Abstract

The invention discloses a photovoltaic cell edge grabbing positioning processing method, which comprises the following steps: a, a distortion correction step: carrying out distortion correction on a camera by using a calibration plate, and generating a position mapping table which is used for recording a corresponding relation between pixel coordinates in a camera view range and world coordinates of the calibration plate; b, an image acquisition and edge extraction step: acquiring an image of the photovoltaic cell, and performing sub-pixel-level edge extraction on the image to obtain pixel point data of the edge of the cell; c, a distortion correction processing step: carrying out distortion correction processing on the edge pixel points based on the position mapping table to obtain corrected edge data; according to the edge-grabbing positioning processing method for the photovoltaic cell, high-precision positioning processing of the photovoltaic cell is realized by using a low-cost wide-angle camera and combining high-precision image processing and geometric transformation technologies.

Description

Technical Field

[0001] The present invention relates to the field of laser processing of photovoltaic cells, and in particular to a method and system for edge grasping and positioning processing of photovoltaic cells. Background Art

[0002] With the rapid development of the photovoltaic industry, the production process of photovoltaic cells has become increasingly complex, and the requirements for processing accuracy have become increasingly higher. Photovoltaic cell laser processing, especially in the manufacture of full back electrode (BC) cells, requires precise laser film opening on the back of the cell to ensure the quality of circuit connection. High-precision processing not only helps to improve product performance, but also reduces material loss and improves production efficiency to a certain extent.

[0003] At present, photovoltaic cell laser film opening technology mostly uses high-precision mechanical positioning or high-cost visual systems, which have problems such as insufficient positioning accuracy or high equipment costs. How to use low-cost hardware (such as wide-angle cameras) combined with efficient image processing algorithms to achieve high-precision positioning to ensure laser processing uniformity is a technical problem that needs to be solved in this field. Summary of the invention

[0004] The present invention aims to provide a high-precision positioning and processing method for photovoltaic cells, using a low-cost wide-angle camera to achieve high-precision (product center position accuracy 0.02mm, angle accuracy 0.01°) positioning and processing. The technical solution of the present invention is as follows: a visual positioning method for laser film opening of a cell, preferably selecting a camera with a suitable field of view based on the size of the cell. The single pixel size is calculated from the camera field of view width W and height H and the camera resolution, and the size of the circular mark engraved on the calibration plate is obtained based on the minimum number of pixels 20 required for feature extraction. The number of circular marks engraved on the calibration plate is obtained from the mark size and the mark spacing.

[0005] Next, place the calibration plate on the processing platform and collect images. Use image region segmentation and feature extraction to identify the approximate position P and radius R of the circular mark on the calibration plate. Generate K calipers based on the contour of the circle around P and R. Calculate the intersection set of the caliper and the circle contour, and fit the pixel coordinates of the center of the circular mark from the intersection set.

[0006] Next, four adjacent circular marks form a rectangular area. The camera's field of view is divided by these rectangular areas. For each area, there are pixel coordinates P1, P2, P3, P4 of the circular mark and the world coordinates W1, W2, W3, W4 of the circular mark on the calibration board. Save these coordinates as a position mapping table.

[0007] Next, the sub-pixel edge extraction method is used to extract the edge data of the four sides of the battery cell.

[0008] Next, according to the pixel position P(x, y) of the edge, find the mapping table of the area where it is located, and use bilinear interpolation to calculate the position after distortion correction.

[0009] Next, use the two-way least squares method to fit the pixel positions of the edge into a straight line with the minimum overall error as the edge of the product.

[0010] Next, the four vertices of the product are obtained by the pairwise intersection of the four sides of the product, and the center position and rotation angle of the product are obtained by the minimum area method.

[0011] Next, use the center position and rotation angle of the product to perform offset and rotation calculations on the processing drawing file to obtain a processing drawing file aligned with the product position.

[0012] Next, parse the processed drawing file after offset and rotation into galvanometer movement and light output commands to process the product.

[0013] The present invention combines methods such as using a reusable high-precision calibration plate to correct the distortion of a wide-angle camera, sub-pixel edge extraction, and two-way least squares straight line fitting, which greatly reduces the hardware cost while ensuring accuracy.

[0014] The present invention provides a method for grasping the edge and positioning the processing of photovoltaic cell wafers. By using a low-cost wide-angle camera and combining high-precision image processing and geometric transformation technologies, high-precision positioning and processing of photovoltaic cell wafers are realized, and the following beneficial effects are obtained: High-precision positioning: By adopting technologies such as sub-pixel edge extraction and two-way least squares straight line fitting, the edge and center position of the photovoltaic cell wafer can be accurately determined, realizing high-precision positioning and processing, and meeting the strict requirements for accuracy in photovoltaic cell production.

[0015] Cost reduction: Using a low-cost wide-angle camera, compared with the traditional high-precision camera vision positioning method, the hardware cost is greatly reduced, and the production efficiency and economic benefits are improved.

[0016] Good stability: By correcting the distortion of the camera and using a position mapping table for distortion correction processing, the stability and accuracy of image processing are improved, ensuring the uniformity of the processing effect and the consistency of battery performance. Description of the Drawings

[0017] Figure 1 : Schematic flow chart of the method for grasping the edge and positioning the processing of photovoltaic cell wafers.

[0018] Figure 2 : Schematic diagram of a caliper on the contour of a circular mark. Detailed Embodiments

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] Please refer to Figure 1 , a method for edge-gripping positioning and processing of photovoltaic cells, comprising the following steps: 1. Distortion correction step 1.1 Placement of calibration plate and image acquisition Place the calibration plate on the processing platform to ensure that its surface is flat and perpendicular to the field of view of the camera. Use the camera to acquire an image of the calibration plate. The calibration plate is marked with a number of circular marks, and the size and arrangement of these circular marks are designed according to the field of view range and resolution of the camera to ensure that the camera can identify at least NxM circular marks.

[0021] 1.2 Image region segmentation and feature extraction Image region segmentation refers to dividing an image into multiple regions for subsequent feature extraction and analysis. In this embodiment, image region segmentation technology is used to divide the image of the calibration plate into multiple regions, and each region contains a circular mark. Feature extraction refers to extracting representative feature information from an image, such as edges, corners, textures, etc. In this embodiment, feature extraction technology is used to identify the approximate position P and radius R of each circular mark. The specific steps are as follows: Grayscale processing: Convert the acquired color image into a grayscale image for subsequent processing and analysis.

[0022] Edge detection: Use an edge detection algorithm (such as the Sobel operator, Canny operator, etc.) to detect the edge information in the image to obtain an edge image.

[0023] Region segmentation: According to the edge image, divide the image of the calibration plate into multiple regions, and each region contains a circular mark.

[0024] Feature extraction: In each region, use a feature extraction algorithm (such as the Hough transform, template matching, etc.) to identify the approximate position P and radius R of the circular mark.

[0025] 1.3 Generation of calipers and fitting of center pixel coordinates As Figure 2 shown, a caliper is a tool for measuring a circle and consists of a set of line segments evenly distributed along the contour of the circle. In this embodiment, according to the approximate position P and radius R of the circular mark, K calipers are generated. The specific steps are as follows: Vernier caliper generation: Centered around the approximate position P of the circular mark, K vernier calipers are evenly generated along the contour of the circle. Each vernier caliper consists of a set of line segments distributed along the contour of the circle, and the lengths and spacings of the line segments are determined according to the radius R of the circular mark.

[0026] Intersection point calculation: Calculate the set of intersection points of each vernier caliper with the contour of the circular mark. Specifically, for each line segment on each vernier caliper, calculate its intersection point with the contour of the circular mark. If the line segment intersects the contour of the circular mark, record the coordinates of the intersection point. Interpolate for all sampling points, find the four pixels closest to the sampling point, and obtain the gray value of the sampling point by distance interpolation. Calculate the average gray value of each set of sampling points, form a gray value curve in the vernier caliper direction, smooth the curve using sigma = 1, find the first derivative of the curve, and determine the extreme points, i.e., the points where the gray value changes the most on the circular contour. Take one point on each side of the extreme point, fit it into a parabola, find the corresponding floating-point coordinates when taking the extreme value, and record them in the point set (Xi, Yi) of the circular contour.

[0027] Fitting of the center pixel coordinates of the circle: According to the set of intersection points, use the least squares method to fit the center pixel coordinates of the circular mark.

[0028] 1.4 Generation of the position mapping table Four adjacent circular marks form a rectangular area. The field of view of the camera is divided by these rectangular areas. For each area, record the pixel coordinates P1, P2, P3, P4 of the circular marks and the world coordinates W1, W2, W3, W4 of the circular marks on the calibration board. Save these coordinates as the position mapping table. The specific steps are as follows: Area division: According to the positions and arrangements of the circular marks, divide the field of view of the camera into multiple rectangular areas. Each rectangular area is composed of four adjacent circular marks.

[0029] Coordinate recording: For each rectangular area, record the pixel coordinates P1, P2, P3, P4 of the circular marks and the world coordinates W1, W2, W3, W4 of the circular marks on the calibration board.

[0030] Position mapping table generation: Save the recorded coordinate information as the position mapping table. The position mapping table is used to record the correspondence between the pixel coordinates and the world coordinates of the calibration board within the field of view of the camera.

[0031] 2. Image acquisition and edge extraction steps 2.1 Image acquisition Use the camera to acquire an image of the photovoltaic cell. The field of view and resolution of the camera are selected according to the size and processing requirements of the photovoltaic cell. The acquired image should contain the complete edge information of the photovoltaic cell.

[0032] 2.2 Pixel-level edge extraction Pixel-level edge extraction refers to the process of extracting edge pixel points from an image. In this embodiment, a grayscale discontinuity detection method is used for pixel-level edge extraction. The specific steps are as follows: Grayscale processing: Convert the captured color image into a grayscale image for subsequent processing and analysis.

[0033] Edge detection: Use edge detection algorithms (such as Sobel operator, Canny operator, etc.) to detect the edge information in the image and obtain an edge image.

[0034] Edge point extraction: Extract the coordinates of the edge points from the edge image to obtain pixel-level edge point data.

[0035] 2.3 Sub-pixel-level edge extraction Sub-pixel-level edge extraction refers to the process of further improving the accuracy of edge points on the basis of pixel-level edge extraction to obtain sub-pixel-level edge points. The edge of the product in the image is composed of a group of connected pixels, which are located at the junction of two regions with discontinuous grayscales. The edge usually shows a jump in grayscale.

[0036] Due to the existence of various noises, in the image captured by the machine vision system, the grayscale change of the pixels at the edge usually is not an ideal step jump, but more tends to be a progressive change. Therefore, a surface model can be used to represent this progressive change.

[0037] In this embodiment, a surface fitting method is used for sub-pixel-level edge extraction. The specific steps are as follows: Grayscale value curve construction: For each pixel-level edge point, construct a grayscale value curve according to its gradient direction. The grayscale value curve represents the grayscale value change along the gradient direction.

[0038] Smoothing processing: Smooth the grayscale value curve to reduce the influence of noise. Smoothing processing can use methods such as moving average filtering and Gaussian filtering.

[0039] Gradient direction calculation: Calculate the gradient direction of the grayscale value curve at each point to determine the gradient direction of the edge point.

[0040] Sub-pixel-level maximum value search: According to the gradient direction, find the sub-pixel-level maximum value on the grayscale value curve. This maximum value is usually located on the straight line where the gradient direction is located.

[0041] Surface fitting: Perform surface fitting on the found sub-pixel-level maximum value, intersect the fitted surface with the straight line where the gradient direction is located, and solve for the extreme value of the intersection point.

[0042] Sub-pixel Edge Coordinate Acquisition: By solving the extreme values of the intersection points, the coordinates of the sub-pixel edge points are obtained. These coordinates are recorded in the edge point set as sub-pixel edge coordinate data.

[0043] 3. Distortion Correction Processing Steps 3.1 Selection of the Mapping Table According to the original position P(x, y) of the edge pixel point, find the mapping table of the area where it is located. The specific method is to determine the rectangular area where it is located according to the coordinates of the edge pixel point, and then select the corresponding mapping relationship from the position mapping table.

[0044] 3.2 Bilinear Interpolation Calculation Use the bilinear interpolation method to calculate the position after distortion correction. The specific steps are as follows: Endpoint Selection: Select four endpoints Q11(x1, y1), Q12(x1, y2), Q21(x2, y1), and Q22(x2, y2) from the mapping table. These four endpoints correspond to the four corner points of the rectangular area where the edge pixel point is located.

[0045] Interpolation in the x-direction: Perform a linear interpolation in the x-direction to obtain two temporary points R1(x, y1) and R2(x, y2). The specific method is to perform a linear interpolation on Q11 and Q21 according to the x coordinate of the edge pixel point to obtain R1; perform a linear interpolation on Q12 and Q22 to obtain R2.

[0046] Interpolation in the y-direction: Perform a linear interpolation in the y-direction to obtain the corrected position Q(x, y). The specific method is to perform a linear interpolation on R1 and R2 according to the y coordinate of the edge pixel point to obtain Q.

[0047] 4. Straight Line Fitting Steps 4.1 Two-way Least Squares Fitting For the list of corrected positions Q1(x, y), Q2(x, y), Q3(x, y)……, use the two-way least squares method to fit a straight line with the minimum overall error. The specific steps are as follows: Assume that the regression line to be found is: y = ax + b. Project all our sample points Q1(x, y), Q2(x, y), Q3(x, y)...... onto the regression line as perpendicular line segments, and use the sum of the squares of these as a metric index to define the objective function:

[0048] The role of the objective function is to quantify the error between the fitted line and the actual data points. By minimizing the objective function, the optimal line parameters a and b can be found so that the fitted line is as close as possible to all data points.

[0049] Parameter Optimization: Solve for the minimum value of the objective function S(a, b) through optimization algorithms (such as gradient descent method, Newton's method, etc.) to obtain the optimal linear parameters a and b. The purpose of solving for the minimum value of the objective function S(a, b) is to find the best linear parameters, minimizing the error between the fitted line and the actual data points, thereby improving the accuracy and reliability of the fitting.

[0050] 5. Steps for Determining the Center Position and Rotation Angle 5.1 Calculation of Vertex Coordinates Calculate the four vertex positions of the product based on the four lines. Specifically, solve for the pairwise intersections of the four lines to obtain the coordinates of the four vertices. Take the arithmetic mean of the vertex coordinates as the center point C.

[0051] 5.2 Determination of the Rotation Angle Using the center point C as the rotation reference, traverse the rotation angles to calculate the area of the circumscribed rectangle. Specifically, rotate the four vertices around the center point C by different angles and calculate the area of the minimum circumscribed rectangle after each rotation. . Take The rotation angle at the minimum is the offset angle of the product.

[0052] 6. Steps for Generating Processing Instructions 6.1 Translation of the Processing Drawing Offset the center coordinates of the processing drawing to the center position coordinates of the photovoltaic cell determined in step 5. Specifically, perform a translation transformation on the processing drawing according to the coordinates of the center point C.

[0053] 6.2 Rotation of the Processing Drawing Rotate the processing drawing according to the rotation angle determined in step 5. Specifically, perform a rotation transformation on the processing drawing according to the rotation angle to align the processing drawing with the actual position of the photovoltaic cell.

[0054] Laser Processing Steps 7.1 Parsing of the Processing Path Perform path parsing on the processed drawing after distortion correction and geometric transformation to generate the processing path. Specifically, generate the path data for laser processing according to the graphic information in the processed drawing.

[0055] 7.2 Control of the Galvo System Calculate the motion trajectory and light emission time parameters of the galvo according to the processing path and send these parameters to the galvo control system to complete the laser processing. Specifically, calculate the motion trajectory and light emission time of the galvo according to the processing path, generate control instructions, send them to the galvo control system, and control the laser to process the photovoltaic cell.

[0056] A system for a method of grasping-edge positioning and processing of a photovoltaic cell, comprising: Camera: used to collect images of photovoltaic cells; Calibration plate: used to correct the distortion of the camera; Image processing module: used to perform sub-pixel edge extraction, distortion correction processing, straight line fitting on the collected images, and calculate the center position and rotation angle of the cell; Processing instruction generation module: used to perform translation and rotation transformation on the processing drawing file according to the center position and rotation angle of the cell, and generate processing instructions; Galvo system: used to receive processing instructions and complete laser processing.

[0057] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A photovoltaic cell edge positioning processing method, characterized in that: The following steps are involved: a. Distortion correction step: Use the calibration plate to correct the camera distortion and generate a position mapping table, which is used to record the correspondence between the pixel coordinates within the camera field of view and the world coordinates of the calibration plate; b. Image acquisition and edge extraction step: Acquire the photovoltaic cell image and perform sub-pixel edge extraction on the image to obtain pixel point data of the cell edge; c. Distortion correction processing step: performing distortion correction processing on edge pixels based on the position mapping table to obtain corrected edge data; d. Linear fitting step: Use the two-way least square method to perform linear fitting on the corrected edge data to obtain a linear equation that characterizes the edge of the battery cell; e. The step of determining the center position and the rotation angle: calculating the intersection of the four sides of the battery cell based on the straight line equation, obtaining the coordinates of the four vertices of the battery cell, and determining the center position and the rotation angle of the battery cell using the minimum area method based on the four vertex coordinates; f. Processing instruction generation step: based on the center position and rotation angle of the battery cell, the processing drawing is translated and rotated to generate a processing instruction aligned with the battery cell position; g. Laser processing steps: parse the processing instructions and control the galvanometer system to complete the laser processing.

2. The method according to claim 1, characterized in that In the distortion correction step a: a1. Calibration board setup: Calculate the single pixel size based on the camera’s field of view and resolution, and determine the diameter of the circular mark and its arrangement on the calibration board; a2. Fitting the pixel coordinates of the center of the circle: Generate multiple calipers evenly along the outline of the circular mark, calculate the intersection set of the calipers and the outline, and fit to obtain the pixel coordinates of the center of the circular mark; a3. Position mapping table generation: Divide the four adjacent circular marks on the calibration plate into rectangular areas, record the mapping relationship between the pixel coordinates of the circular marks in each area and the world coordinates of the calibration plate, and generate a position mapping table.

3. The method according to claim 1, characterized in that In the image acquisition and edge extraction step b: b1. Image preprocessing: preprocessing and noise smoothing of the collected images; b2. Preliminary edge detection: Use gray discontinuity to detect the preliminary edge position and construct a gray value curve based on the gradient direction; b3. Obtaining sub-pixel edge coordinates: Smooth the grayscale value curve, and find the maximum value at the sub-pixel level through the gradient direction of the edge point in the gradient image; use the surface fitting method to intersect the fitted surface with the straight line where the gradient direction is located, solve the extreme value of the intersection, and obtain the sub-pixel edge coordinate data.

4. The method according to claim 1, characterized in that In the distortion correction processing step c: c1. Mapping table selection: According to the original position P(x, y) of the edge pixel and the mapping table corresponding to its area, select the four endpoints Q11(x1, y1), Q12(x1, y2), Q21(x2, y1) and Q22(x2, y2) in the mapping table; c2. Bilinear interpolation calculation: Using the bilinear interpolation method, first perform a linear interpolation in the x direction to obtain the intermediate value, and then interpolate in the y direction to calculate the corrected pixel position Q(x,y).

5. The method according to claim 1, characterized in that In the straight line fitting step d: d1. Bidirectional least squares fitting: Perform bidirectional least squares fitting on all sampling points in the corrected edge data, considering the horizontal and vertical errors respectively; d2. Objective function optimization: The sum of the squares of the vertical distances from all sample points to the fitting line is used as the objective function for optimization. When the objective function reaches the minimum value, the corresponding straight line parameters are used as the best fitting straight line parameters of the battery cell edge.

6. The method according to claim 1, characterized in that In the center position and rotation angle determination step e: e1. Vertex coordinate calculation: Calculate the four vertex positions of the product based on the four straight lines, and take the arithmetic average of the vertex coordinates as the center point; e2. Determine the rotation angle: Take the center point as the rotation reference, traverse the rotation angles to calculate the area of ​​the circumscribed rectangle, and select the rotation angle with the smallest area as the final offset angle.

7. The method according to claim 1, characterized in that In the processing instruction generation step f: f1. Translation of the processing drawing: offset the center coordinates of the processing drawing to the center position coordinates of the photovoltaic cell determined in step e; f2. Processing drawing rotation: Rotate the processing drawing according to the rotation angle determined in step e, so that the processing drawing is aligned with the actual position of the photovoltaic cell.

8. The method according to claim 1, characterized in that In the laser processing step g: g1. Processing path analysis: Analyze the processing image after distortion correction and geometric transformation to generate the processing path; g2. Galvanometer system control: Calculate the motion trajectory and light emission time parameters of the galvanometer according to the processing path, and send these parameters to the galvanometer control system to complete the laser processing.

9. A system for the photovoltaic cell edge grasping and positioning processing method according to any one of claims 1 to 8, characterized in that: include: Camera: used to collect images of photovoltaic cells; Calibration board: used to correct camera distortion; Image processing module: used to perform sub-pixel edge extraction, distortion correction, straight line fitting, and calculation of the center position and rotation angle of the battery cell on the collected images; Processing instruction generation module: used to translate and rotate the processing drawing according to the center position and rotation angle of the battery cell to generate processing instructions; Galvanometer system: used to receive processing instructions and complete laser processing.

Images