Photovoltaic cell production positioning method and system based on image processing

Abstract

The invention relates to the field of image processing, in particular to a photovoltaic cell production positioning method and system based on image processing, and the method comprises the steps: obtaining a photovoltaic cell image, calculating a corner response function value of each pixel point in the image through employing a Harris algorithm, taking the pixel point with the corner function value greater than an optimal threshold value as a real corner point, and carrying out the calculation of the corner response function value of each pixel point; and determining the position of the photovoltaic cell through the real angular point. According to the method, the accuracy of the threshold is improved, so that the real angular point can be obtained, and the photovoltaic cell can be conveniently positioned.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a photovoltaic cell production positioning method and system based on image processing. Background Art

[0002] In the production process of photovoltaic cells, from the selection of raw materials, the processing of wafers, to the packaging of photovoltaic modules, precise positioning is required to ensure the quality and efficiency of the product. Specifically, the initial process usually starts with the cutting of silicon wafers, which requires precise positioning of the silicon wafers to ensure that the size of each wafer meets the requirements. After the silicon wafers are cut into cells, the cells need to be connected into strings in a certain order. The key to this process is to accurately position the cells to ensure that they remain in the correct position during wiring and welding. After the photovoltaic cells are connected in series and wired, the cells need to be laminated with other components (such as backplanes, transparent glass, etc.) and finally packaged into complete photovoltaic panels. In the production process of photovoltaic panels, precise positioning is essential to ensure the quality and efficiency of the product.

[0003] The Chinese patent application document with publication number CN108629343A discloses a license plate positioning method and system based on edge detection and Harris corner point detection. The method includes: S1, inputting an image, performing smoothing filtering on the image, and obtaining a smoothed image; S2, performing edge detection on the smoothed image, and obtaining an edge detection map; S3, performing corner point detection on the edge detection map, obtaining the corner point position, and completing the license plate positioning according to the corner point position.

[0004] In the related art, the Harris algorithm can be used to detect corner points on photovoltaic cells, and the position of the photovoltaic cells can be determined by the corner points, thereby realizing the positioning of the photovoltaic cells. However, the Harris algorithm in the related art determines whether a pixel is a corner point by comparing the response function value of the pixel point with a threshold, and the threshold value is usually set based on experience, and the accuracy of the threshold value is poor, which ultimately leads to low accuracy in corner point detection, affecting the positioning of the photovoltaic panel. Summary of the invention

[0005] In order to improve the accuracy of threshold setting, the present invention provides a photovoltaic cell production positioning method and system based on image processing.

[0006] In a first aspect, the present invention provides a photovoltaic cell production positioning method based on image processing, which adopts the following technical solution: Obtain a photovoltaic cell image, use the Harris algorithm to calculate the corner point response function value of each pixel in the image, take the pixel whose corner point function value is greater than the optimal threshold as the real corner point, and determine the position of the photovoltaic cell through the real corner point; The optimal threshold is calculated as follows: An initial threshold is set, and pixels whose corner point response function values ​​are greater than the initial threshold are taken as pre-selected corner points. The gradient continuity of the pre-selected corner points in each direction is calculated to obtain multiple gradient continuity values, which represent the gradient variation characteristics of the neighboring pixel points in the corresponding direction. The product of the maximum gradient continuity and the corner point response function value is taken as the effective eigenvalue of the pre-selected corner point. The pre-selected corner points whose effective eigenvalues ​​are greater than the initial threshold are classified as the first category, and those whose effective eigenvalues ​​are greater than the initial threshold are classified as the second category. Construct the objective function: ; In the formula, is the objective function, is the intra-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the first class, is the intra-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the second class, is the inter-class variance of the effective eigenvalues ​​of the pre-selected corner points in the first category and the effective eigenvalues ​​of the pre-selected corner points in the second category; the initial threshold when the objective function is the minimum value is taken as the optimal threshold.

[0007] The effect is that the optimal threshold is obtained by calculation, which improves the accuracy of the threshold. The optimal threshold can accurately screen out noise points and real corner points, and further judge the position of the photovoltaic cell by the real corner points, thereby improving the accuracy of photovoltaic cell positioning.

[0008] Preferably, the gradient continuity is calculated as follows: A window is constructed with the pre-selected corner point as the center, and the gradient continuity in each direction within the window area is calculated. The expression of gradient continuity is: ; In the formula, is the gradient continuity of the preselected corner point a in the β direction, is the mean value of the gradient amplitude of the pixels in the β direction in the neighborhood of the pre-selected corner point a, is the variance of the gradient amplitude of the pixel points in the β direction in the neighborhood of the pre-selected corner point a, It is the gradient direction consistency of the pixels in the β direction within the neighborhood of the pre-selected corner point a.

[0009] The effect is that the gradient continuity is obtained by calculating multiple dimensions. The gradient continuity can be used to preliminarily identify noise points and real corner points, thereby improving the accuracy of screening noise points.

[0010] Preferably, the gradient continuity is calculated as follows: A window is constructed with the pre-selected corner point as the center, and the gradient continuity in each direction within the window area is calculated. The expression of gradient continuity is: ; In the formula, is the gradient continuity of the preselected corner point a in the β direction, is the mean value of the gradient amplitude of the pixels in the β direction in the neighborhood of the pre-selected corner point a, is the variance of the gradient amplitude of the pixel points in the β direction in the neighborhood of the pre-selected corner point a, It is the gradient direction consistency of the pixels in the β direction within the neighborhood of the pre-selected corner point a.

[0011] Preferably, the gradient direction consistency calculation method of the pixel points in the β direction is: The cosine similarity between the gradient vector and the mean vector of each pixel is calculated, and the average value of multiple cosine similarities is taken as the gradient direction consistency.

[0012] The effect is that by calculating the consistency, the possibility of the corresponding pixel point being an edge line can be determined, which facilitates further screening of noise points and real corner points.

[0013] Preferably, the gradient direction consistency calculation method of the pixel points in the β direction is: Calculate the gradient direction of each pixel and take the majority of the gradient directions as the reference direction; A direction interval is constructed with the reference direction as the center, and the pixels whose gradient directions are within the direction interval are taken as standard pixels. The ratio of the number of standard pixels to the total number of pixels in the β direction is taken as the gradient direction consistency.

[0014] Preferably, the Sobel operator is used to calculate the gradient amplitude and gradient direction of the pixel point.

[0015] The effect is that by calculating the gradient magnitude and gradient direction, it is easier to calculate the gradient continuity.

[0016] Preferably, a swarm intelligence algorithm is used to solve the minimum value of the objective function to obtain the optimal threshold.

[0017] The effect is that the optimal threshold is obtained by solving the objective function, and the real corner point can be judged by using the optimal threshold, which is convenient for positioning the photovoltaic cell.

[0018] In a second aspect, the present invention provides a photovoltaic cell production positioning system based on image processing, which adopts the following technical solutions: A photovoltaic cell production positioning system based on image processing, a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the photovoltaic cell production positioning method based on image processing is implemented.

[0019] The effect is that the above-mentioned photovoltaic cell production positioning method based on image processing is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor for easy use.

[0020] The present invention has the following technical effects: By establishing an objective function and solving the objective function to obtain an optimal threshold, compared with the manually set threshold in the prior art, the optimal threshold calculated by the present invention improves the accuracy of the threshold; the optimal threshold can be used to screen out noise points and real corner points, and the position of the photovoltaic cell can be further determined by the real corner points, thereby improving the accuracy of photovoltaic cell positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0022] Figure 1 It is a flow chart of the photovoltaic cell production positioning method based on image processing of the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0025] The embodiment of the present invention discloses a photovoltaic cell production positioning method based on image processing, referring to Figure 1 , including the following steps, as follows: S1: Obtain a photovoltaic cell image and use the Harris algorithm to calculate the corner point response function value of each pixel in the image.

[0026] In the production process of photovoltaic cells, a camera is used to collect images of the photovoltaic cells to be positioned, and the collected photovoltaic cell images are converted into grayscale images. The sobel operator is used to calculate the gradient amplitude and gradient direction of the pixel points in the grayscale image, and then the Harris algorithm is used to calculate the corner point response function value of each pixel point in the image.

[0027] S2: Calculate the optimal threshold.

[0028] S21: Setting an initial threshold, and taking pixels whose corner point response function values ​​are greater than the initial threshold as pre-selected corner points.

[0029] After obtaining the corner point response function value of each pixel in the image, an initial threshold is set artificially to delete the pixels with low response function values ​​that cannot be corner points, and retain the pixels that may be corner points as pre-selected corner points. The pre-selected corner points will contain some noise points.

[0030] Exemplarily, the initial threshold is set to 1. When the corner point response function value of a pixel point does not exceed 1, the corresponding pixel point is not a corner point and is deleted. Otherwise, the point is retained as a pre-selected corner point. In actual applications, the gradient changes of a few noise points are large, resulting in large response function values, which cannot be accurately distinguished from real corner points. It should be noted that the initial threshold is a hyperparameter, and a lower threshold should be selected, which is only used to screen out pixel points that cannot be corner points.

[0031] S22: Calculate the gradient continuity of the pre-selected corner point in each direction to obtain multiple gradient continuities, where the gradient continuity represents the change characteristics of the gradient of the neighboring pixel points in the corresponding direction.

[0032] In one embodiment, a window is constructed with a preselected corner point as the center, the window size is 9×9, and the preselected corner point is used as the origin. Eight rays are uniformly constructed, each ray represents a direction, and there are eight directions in total. The gradient continuity in each direction in the window area is calculated, and a total of 8 gradient continuities are obtained.

[0033] The expression for gradient continuity is: ; In the formula, is the gradient continuity of the preselected corner point a in the β direction, is the mean value of the gradient amplitude of the pixels in the β direction in the neighborhood of the pre-selected corner point a, is the variance of the gradient amplitude of the pixel points in the β direction in the neighborhood of the pre-selected corner point a, The consistency of the gradient direction of the pixels in the β direction of the pre-selected corner point a in the neighborhood is calculated by calculating the cosine similarity between the gradient vector of each pixel and the mean vector, and taking the average of multiple cosine similarities as the gradient direction consistency.

[0034] The larger the mean value of the gradient amplitude is, the larger the gradient amplitude is in that direction, the more dramatic the change in the grayscale value in the perpendicular direction of the β direction is, and the greater the possibility that the pixel point in the β direction is an edge line; the smaller the variance of the pre-selected corner point a in the β direction is, the smaller the gradient change in that direction is, and the greater the continuity is; the greater the consistency is, the more similar the gradient vector of the pixel point is to the mean vector; in summary, the larger the value of the gradient continuity in the β direction is, the greater the possibility that the pixel point in the β direction is an edge line, the greater the distinction between the pre-selected corner point a and the noise point is, and further indicates that the pre-selected corner point a is more likely to be the true angle.

[0035] In one embodiment, a method for calculating the gradient direction consistency of pixel points in the β direction is as follows: calculate the gradient direction of each pixel point, and use the majority of the gradient directions as a reference direction; construct a direction interval with the reference direction as the center, and use the pixel points whose gradient directions are within the direction interval as standard pixel points, and use the ratio of the number of standard pixel points to the total number of pixel points in the β direction as the gradient direction consistency.

[0036] For example, among the 8 pixels in the 0 degree direction, the mode of the gradient direction is 85 degrees, and the number of modes is 5. A direction interval (85-5, 85+5) is constructed with 85 degrees as the center, that is, the interval is (80, 90), and the unit is: degree. The pixel points whose gradient direction is within the interval are taken as standard pixel points. Assuming that the number of standard pixel points is 7, the consistency of the gradient direction is 7 / 8=0.875.

[0037] In one embodiment, the gradient continuity is calculated as: A window is constructed with the pre-selected corner point as the center, and the gradient continuity in each direction within the window area is calculated. The expression of gradient continuity is: ; In the formula, is the gradient continuity of the preselected corner point a in the β direction, is the mean value of the gradient amplitude of the pixels in the β direction in the neighborhood of the pre-selected corner point a, is the variance of the gradient amplitude of the pixel points in the β direction in the neighborhood of the pre-selected corner point a, It is the gradient direction consistency of the pixels in the β direction within the neighborhood of the pre-selected corner point a.

[0038] S23: taking the product of the maximum gradient continuity and the corner point response function value as the effective eigenvalue of the pre-selected corner point, and classifying the pre-selected corner points whose effective eigenvalues ​​are greater than the initial threshold as the first category, and vice versa as the second category.

[0039] The expression for the effective eigenvalue is: ; In the formula, To preselect corner point a as a valid eigenvalue, is the corner point response function value of the preselected corner point a, is the maximum value of the gradient continuity of the pre-selected corner point a in eight directions.

[0040] The larger the value, the greater the distinction between the pre-selected corner point a and other points. The larger the effective eigenvalue is, the greater the distinction between the pre-selected corner point a and the noise point is. In summary, the larger the effective eigenvalue is, the greater the possibility that the pre-selected corner point a is a real corner point is, and the noise point can be effectively screened out through the effective eigenvalue.

[0041] The pre-selected corner points are classified using the initial threshold and the effective eigenvalue, wherein the pre-selected corner points whose effective eigenvalues ​​are greater than the initial threshold are classified into the first category, and the pre-selected corner points whose effective eigenvalues ​​are less than or equal to the initial threshold are classified into the second category, and the intra-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the first category and the second category are calculated, and the inter-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the first category and the effective eigenvalues ​​of the pre-selected corner points in the second category are calculated, wherein the calculation method of the intra-class variance and the inter-class variance is the existing technology, and the specific process will not be repeated here.

[0042] S24: construct an objective function, and solve the objective function to obtain an optimal threshold.

[0043] The expression of the objective function is: ; In the formula, is the objective function, is the intra-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the first class, is the intra-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the second class, is the inter-class variance of the validity eigenvalues ​​of the pre-selected corner points in the first category and the valid eigenvalues ​​of the pre-selected corner points in the second category.

[0044] The minimum value of the objective function is solved by using a swarm intelligence algorithm, and the initial threshold when the objective function is the minimum value is taken as the optimal threshold. The swarm intelligence algorithm is any one of the particle swarm algorithm, ant colony algorithm, artificial bee colony algorithm, artificial fish swarm algorithm, and firefly algorithm.

[0045] S3: Pixel points whose corner point function values ​​are greater than the optimal threshold are taken as true corner points, and the positions of the photovoltaic cells are determined through the true corner points.

[0046] The optimal threshold can be used to filter out noise points and real corner points, where pixels greater than the optimal threshold are real corner points, and pixels less than or equal to the optimal threshold are noise points. In the actual production process of photovoltaic cells, the position of the photovoltaic cell is determined by the real corner point, so as to adjust the position of the photovoltaic cell to achieve the positioning of the photovoltaic cell.

[0047] An embodiment of the present invention also discloses a photovoltaic cell production positioning system based on image processing, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a photovoltaic cell production positioning method based on image processing according to the present invention is implemented.

[0048] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0049] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high bandwidth memory HBM (High Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0050] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0051] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A photovoltaic cell production positioning method based on image processing, characterized in that: Includes steps: Obtain a photovoltaic cell image, use the Harris algorithm to calculate the corner point response function value of each pixel in the image, take the pixel whose corner point function value is greater than the optimal threshold as the real corner point, and determine the position of the photovoltaic cell through the real corner point; The optimal threshold is calculated as follows: An initial threshold is set, and pixels whose corner point response function values ​​are greater than the initial threshold are used as pre-selected corner points. The gradient continuity of the pre-selected corner points in various directions is calculated to obtain multiple gradient continuity values. The gradient continuity represents the change characteristics of the gradient of the neighboring pixel points in the corresponding direction. The product of the maximum gradient continuity and the corner point response function value is used as the effective eigenvalue of the pre-selected corner point; the pre-selected corner points whose effective eigenvalues ​​are greater than the initial threshold are classified as the first category, and vice versa as the second category; Construct the objective function: ; In the formula, is the objective function, is the intra-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the first class, is the intra-class variance of the effectiveness eigenvalues ​​of the pre-selected corner points in the second class, is the inter-class variance of the effective eigenvalues ​​of the pre-selected corner points in the first category and the effective eigenvalues ​​of the pre-selected corner points in the second category; the initial threshold when the objective function is the minimum value is taken as the optimal threshold.

2. The photovoltaic cell production positioning method based on image processing according to claim 1, characterized in that: The gradient continuity is calculated as: A window is constructed with the pre-selected corner point as the center, and the gradient continuity in each direction within the window area is calculated. The expression of gradient continuity is: ； In the formula, is the gradient continuity of the preselected corner point a in the β direction, is the mean value of the gradient amplitude of the pixels in the β direction in the neighborhood of the pre-selected corner point a, is the variance of the gradient amplitude of the pixel points in the β direction in the neighborhood of the pre-selected corner point a, It is the gradient direction consistency of the pixels in the β direction within the neighborhood of the pre-selected corner point a.

3. The photovoltaic cell production positioning method based on image processing according to claim 1, characterized in that: The gradient continuity is calculated as: A window is constructed with the pre-selected corner point as the center, and the gradient continuity in each direction within the window area is calculated. The expression of gradient continuity is: ； In the formula, is the gradient continuity of the preselected corner point a in the β direction, is the mean value of the gradient amplitude of the pixels in the β direction in the neighborhood of the pre-selected corner point a, is the variance of the gradient amplitude of the pixel points in the β direction in the neighborhood of the pre-selected corner point a, It is the gradient direction consistency of the pixels in the β direction within the neighborhood of the pre-selected corner point a.

4. The photovoltaic cell production positioning method based on image processing according to claim 2 or 3, characterized in that: The calculation method of the gradient direction consistency of the pixel point in the β direction is: The cosine similarity between the gradient vector and the mean vector of each pixel is calculated, and the average value of multiple cosine similarities is taken as the gradient direction consistency.

5. The photovoltaic cell production positioning method based on image processing according to claim 2 or 3, characterized in that: The calculation method of the gradient direction consistency of the pixel point in the β direction is: Calculate the gradient direction of each pixel and take the majority of the gradient directions as the reference direction; A direction interval is constructed with the reference direction as the center, and the pixels whose gradient directions are within the direction interval are taken as standard pixels. The ratio of the number of standard pixels to the total number of pixels in the β direction is taken as the gradient direction consistency.

6. The photovoltaic cell production positioning method based on image processing according to any one of claims 1 to 3, characterized in that: The Sobel operator is used to calculate the gradient magnitude and gradient direction of the pixel.

7. The photovoltaic cell production positioning method based on image processing according to claim 1, characterized in that: The swarm intelligence algorithm is used to solve the minimum value of the objective function and obtain the optimal threshold.

8. Photovoltaic cell production positioning system based on image processing, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the photovoltaic cell production positioning method based on image processing according to any one of claims 1 to 7 is implemented.

Images