A PECVD Edge Plating Identification Method and System Based on HSV Model

Abstract

The present invention discloses a method and system for identifying PECVD overcoating based on the HSV model, which relates to the technical field of overcoating identification and solves the problems of low efficiency and low accuracy in existing PECVD overcoating identification. The key points of the solution are as follows: obtaining an image of a battery wafer after coating collected by a CCD camera; extracting the image of the card position point from the battery wafer image; converting the image of the card position point into the HSV color space to obtain the HSV channel values of the image of the card position point, and graying the image of the card position point to obtain the gray value of the image of the card position point; comparing the HSV channel values and gray values of the image of the card position point with a set range to determine whether there is a card position point, and comparing the number of card position points with the actual number of card position points of the battery wafer to determine whether there is overcoating on the battery wafer; identifying whether the card position points of the battery wafer are normal through the combination of the HSV channel values and gray values of the image of the card position point in the battery wafer image, and judging whether there is an overcoating situation on the battery wafer in combination with the number of normal card position points.

Description

Technical Field

[0001] The present invention relates to the technical field of plating-around recognition, and more specifically, it relates to a method and system for PECVD plating-around recognition based on the HSV model. Background Art

[0002] PECVD (Plasma Enhanced Chemical Vapor Deposition) is a technology used to deposit thin films on the surface of silicon wafers and is widely applied in the production of solar cells. In the production of TOPCon solar cells, the PECVD technology can be used to deposit a passivation layer and a polysilicon layer. Plating-around means that during the deposition process, the thin film is formed not only on the front side of the silicon wafer but also on the edges and the back side of the silicon wafer, which may affect the quality and electrical performance of the battery. Therefore, PECVD plating-around recognition is an issue that needs attention.

[0003] Traditional PECVD plating-around recognition methods mainly rely on pure visual inspection, where the edges and the back side of the silicon wafer are observed through a microscope or the naked eye to check for unnecessary deposits. The traditional pure visual inspection method has problems such as low efficiency and low accuracy, and is prone to missed judgments, misjudgments, or overjudgments.

[0004] Therefore, the present application provides a method and system for PECVD plating-around recognition based on the HSV model to solve the above problems. Summary of the Invention

[0005] The objective of the present application is to provide a method and system for PECVD plating-around recognition based on the HSV model to solve the problems of low efficiency and low accuracy in existing PECVD plating-around recognition; by extracting the HSV channel values and grayscale values of the card positions of the battery cell to determine the number of card positions, and then identifying the plating-around battery cell to solve the above problems.

[0006] The present application first provides a method for PECVD plating-around recognition based on the HSV model, including: obtaining an image of the battery cell after coating collected by a CCD camera; extracting all card position images from the battery cell image; converting the card position image to the HSV color space to obtain the HSV channel values of the card position image, and gray-scaling the card position image to obtain the grayscale value of the card position image; comparing the HSV channel values and grayscale values of the card position image with a set range to determine whether there are card positions, and comparing the number of card positions with the actual number of card positions of the battery cell to determine whether the battery cell has plating-around.

[0007] By adopting the above technical solution, by jointly identifying whether there is a plating-around situation of the battery cell by extracting the HSV channel values and grayscale values of the card position image of the battery cell, it is more accurate and faster compared with the traditional pure visual inspection, and has a lower error rate compared with the RGB color space recognition.

[0008] In a possible implementation, a setting range is obtained through the following steps: acquiring cell images with the film thickness at the normal upper limit, the film thickness at the normal lower limit, and the module at the normal median value collected by a standard-mode CCD camera; extracting the card point position images from the cell images, and measuring the HSV channel values and grayscale values of the card point position images; calculating the average HSV channel value and the average grayscale value of all the card point position images as the reference values; setting a deviation value centered on the reference values to obtain the HSV channel setting range and the grayscale value setting range.

[0009] In a possible implementation, comparing the HSV channel values and grayscale values of the card point position images with the setting range to determine whether there are card point positions, and comparing the number of card point positions with the actual number of card point positions of the cell to determine whether there is plating around the cell, including: comparing the HSV channel values and grayscale values of the card point position images with the setting range, and when both the HSV channel values and grayscale values of the card point position images are within the setting range, determining that there are card point positions and marking them as 1, otherwise marking them as 0; counting the sum of the marks of all the card point images on the same cell image, and determining it as qualified when the sum of the marks is equal to the actual number of card point positions of the cell, otherwise determining it as unqualified.

[0010] In a possible implementation, extracting all the card point position images from the cell images, including: extracting the cell coordinate data from the cell image based on the AOI software; extracting the rectangular frame coordinates centered on the card point positions based on the cell coordinate data; extracting the card point position images from the cell image according to the rectangular frame coordinates.

[0011] In a possible implementation, the setting range is: H channel value 170.8 ± 15, S channel value 64.2 ± 10, V channel value 46.9 ± 10, grayscale value 25 ± 5.

[0012] The present application also provides a PECVD plating-around recognition system based on the HSV model, including: a data acquisition unit for acquiring cell images after film plating collected by a CCD camera; a card point position extraction unit for extracting all the card point position images from the cell images; a determination data extraction unit for converting the card point position images into the HSV color space to obtain the HSV channel values of the card point position images, and gray-scaling the card point position images to obtain the grayscale values of the card point position images; a plating-around determination unit for comparing the HSV channel values and grayscale values of the card point position images with the setting range to determine whether there are card point positions, and comparing the number of card point positions with the actual number of card point positions of the cell to determine whether there is plating around the cell.

[0013] In a possible implementation, the plating-around determination unit includes: a setting range extraction unit configured to obtain cell images with film thickness at the normal upper limit, film thickness at the normal lower limit, and the module at the normal median collected by a standard-mode CCD camera; extract the card point position images from the cell images, and measure the HSV channel values and grayscale values of the card point position images; calculate the average HSV channel value and the average grayscale value of all the card point position images as the reference values; set a deviation value centered on the reference values to obtain the HSV channel setting range and the grayscale value setting range.

[0014] In a possible implementation, the plating-around determination unit is specifically configured to compare the HSV channel values and grayscale values of the card point position images with the setting range. When both the HSV channel values and grayscale values of the card point position images are within the setting range, it is determined that there is a card point position, which is marked as 1; otherwise, it is marked as 0. The sum of the marks of all the card point images on the same cell image is statistically calculated. When the sum of the marks is equal to the actual number of card point positions of the cell, it is determined to be qualified; otherwise, it is determined to be unqualified.

[0015] In a possible implementation, the card point position extraction unit is specifically configured to extract the cell coordinate data from the cell image based on the AOI software; extract the rectangular frame coordinates centered on the card point position based on the cell coordinate data; and extract the card point position images from the cell image according to the rectangular frame coordinates.

[0016] In a possible implementation, the setting range in the plating-around determination unit is: H channel value 170.8 ± 15, S channel value 64.2 ± 10, V channel value 46.9 ± 10, grayscale value 25 ± 5.

[0017] Compared with the prior art, the present application has the following beneficial effects: By identifying whether it is normal in the card point position image through the setting range, and then analyzing whether there is plating around the cell by the number of normal card point positions and the actual number of card point positions of the cell, it is more accurate and faster compared with the traditional pure vision detection; By extracting the HSV channel values and cooperating with the grayscale values to judge the plating around the card point position, the HSV color space can not only identify the edge film color change, but also identify the hue and saturation changes, enhancing the accuracy of plating-around identification. Compared with the RGB color space identification, the error rate is lower. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not constitute a limitation to the embodiments of the present invention. In the drawings:

[0019] Figure 1 is a schematic diagram of the card point position;

[0020] Figure 2 is a flowchart of the PECVD plating-around identification method based on the HSV model;

[0021] Figure 3 Schematic diagram of device interaction for PECVD overcoating identification method;

[0022] Figure 4 Structural diagram of PECVD overcoating identification system based on HSV model. Specific implementation manners

[0023] To make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below in conjunction with embodiments and the accompanying drawings. The illustrative embodiments and descriptions of the present application are only used to explain the present application and shall not be construed as a limitation to the present application.

[0024] First, the terms in the present application will be described to facilitate understanding of the technical content of the present application.

[0025] Card point position: It refers to the position where the mechanical structures for fixing the battery cells (silicon wafers) are located in the PECVD equipment. These mechanical structures ensure that the silicon wafers maintain the correct position during the coating process. In a general PECVD equipment, a single silicon wafer is fixed by three card points, and the corresponding card point positions are as shown in the triangular area in Figure 1 as shown.

[0026] Please refer to Figure 2 as shown, Figure 2 Flowchart of the PECVD overcoating identification method based on the HSV model. The method includes: S1, obtaining the image of the battery cell after coating collected by the CCD camera; S2, extracting all card point position images from the battery cell image; S3, converting the card point position image to the HSV color space to obtain the HSV channel values of the card point position image, and graying the card point position image to obtain the gray values of the card point position image; S4, comparing the HSV channel values and gray values of the card point position image with the set range to determine whether there are card point positions, and comparing the number of card point positions with the actual number of card point positions of the battery cell to determine whether the battery cell has overcoating.

[0027] The principle of realizing PECVD overcoating identification by this method is that under normal circumstances, due to the card point positions being blocked, the color of the card point positions approaches that of the original silicon; when overcoating occurs, there is additional deposition at the card point positions, and the color of the card point positions is different from that of the original silicon. Therefore, it is possible to accurately judge whether the battery cell has overcoating by comparing the color difference between the card point positions and the normal coating area after the battery cell is coated.

[0028] Specifically, before executing this method, first place a single battery cell accurately on the carrying graphite boat to ensure that the three card points are in firm contact with the battery cell, and at the same time check the contact situation between the card points and the battery cell to ensure that its position is accurate and correct, and then perform coating on the battery cell. Then, as shown in Figure 3As shown, a CCD camera with adjusted parameters in the automatic optical inspection (AOI, Automated Optical Inspection) system is used to collect a two-dimensional rectangular image of the cell, and the collected cell image is transmitted to an electronic computer for analysis. During the analysis, the HSV channel value and grayscale value of the card point image are extracted to identify the number of card points, and the cell is judged whether there is bypass plating based on the number of identified card points. If the number of identified card points is less than three, it is judged as an NG bypass plated cell; if the number of identified card points is three, it is judged as an OK normal cell.

[0029] It should be noted that the inventors found that the number of card points for PECVD wrap-around cell plates is lower than that for normal cell plates, and there are uncoated areas at the card points due to occlusion, and the color of this area is different from that of the normal area. Therefore, the inventors identify whether the card point image is normal by setting a range, and then analyze whether the cell has wrap-around plating by the normal number of card points and the actual number of card points for the cell. In addition, the inventors also found that the color change of the wrap-around edge is narrow and the transition is soft, and the numerical change in the RGB image is small. When the wrap-around judgment of the card point is based on the RGB color space, it is easy to miss the wrong judgment. Therefore, the inventors extract the HSV channel value and the grayscale value to judge the wrap-around plating of the card point. The HSV color space can not only identify the edge film color change, but also the hue and saturation changes, and enhance the accuracy of wrap-around identification.

[0030] It can be seen that the improvement of this solution is that by extracting the HSV channel value and grayscale value of the cell card point image, it is jointly identified whether the cell has bypass plating. Compared with traditional pure visual inspection, it is more accurate and faster, and compared with RGB color space recognition, it has a lower error rate.

[0031] Step S1 is to obtain the image of the cell after coating by the CCD camera. Specifically, the parameters of the CCD camera in the AOI system can be adjusted to collect high-definition images; a cell (182.2mm*183.75mm) with N-type coating and three card points is placed under the CCD camera to capture the cell image.

[0032] Step S2 is to extract all the card point images from the cell image. Specifically, it includes: extracting the cell coordinate data from the cell image based on the positioning function of the AOI software, and performing rectangular coordinate selection based on the cell coordinate data and the coordinates of the three preset card points. A 2mm 2 The rectangular coordinate frame (the groove depth of the card point is 0.9mm, with a small offset up and down) is obtained, and the rectangular frame coordinates centered on the three card points are obtained, and the three card point images are extracted according to the rectangular frame coordinates.

[0033] Step S3 is to convert the card position image to the HSV color space to obtain the HSV channel values of the card position image, and to grayscale the card position image to obtain the grayscale values of the card position image. Specifically, the image collected by the CCD camera is in RGB mode and needs to be converted to HSV mode for processing. Also, in order to obtain a higher-quality grayscale image, the color also needs to be converted to the grayscale space.

[0034] First, use image software to extract the RGB three-channel values of the selected card position image to obtain the RGB data of the card position image, and calculate the grayscale values and HSV channel values according to the RGB data.

[0035] Convert the RGB color values to HSV color values according to a specific formula; the algorithm for converting RGB to HSV is as follows:

[0036] Let max = max(R, G, B), min = min(R, G, B);

[0037] V = max(R, G, B);

[0038] S = (max - min) / max;

[0039] If R = max, H = (G - B) / (max - min)*60;

[0040] If G = max, H = 120 + (B - R) / (max - min)*60;

[0041] If B = max, H = 240 + (R - G) / (max - min)*60;

[0042] If H < 0, H = H + 360.

[0043] It is also possible to directly convert the code using functions in existing image processing libraries, such as the OpenCV or PIL (Pillow) libraries.

[0044] Secondly, calculate the corresponding grayscale values according to the common formula for converting RGB to a grayscale image. The common formula for converting RGB to a grayscale image is: gray = r*0.299 + g*0.587 + b*0.114, where r, g, and b represent the values of the three RGB channels respectively.

[0045] Step S4 is to compare the HSV channel values and grayscale values of the card position images with the set ranges to determine whether there are card positions, and to compare the number of card positions with the actual number of card positions of the solar cell to determine whether there is overplating on the solar cell. Specifically, it includes: comparing the HSV channel values and grayscale values of the card position images with the set ranges. When both the HSV channel values and grayscale values of the card position images are within the set ranges, it is determined that there are card positions, and it is marked as 1; otherwise, it is marked as 0. Count the sum of the marks of all card position images on the same solar cell image. When the sum of the marks is equal to the actual number of card positions of the solar cell, it is determined to be qualified; otherwise, it is determined to be unqualified.

[0046] Specifically, when the hue (H), saturation (S), value (V), and grayscale value (gray) of the card position image are within the set ranges, the identification number 1 will be marked; otherwise, 0 will be marked. The sum of the marks of the three card position images of the same solar cell is counted by software. If the sum is equal to 3, it is determined as an OK normal solar cell; if it is less than 3, it is determined as an NG overplated solar cell.

[0047] The set ranges in step S4 are obtained through the following steps: Obtain the solar cell images with the film thickness at the normal upper limit, the film thickness at the normal lower limit, and the module at the normal median collected by the standard mode CCD camera; Extract the card position images from the solar cell images, and measure the HSV channel values and grayscale values of the card position images; Calculate the average values of the HSV channel values and grayscale values of all card position images as the reference values; Set the deviation values centered on the reference values to obtain the HSV channel set ranges and grayscale value set ranges.

[0048] Specifically, adjust the CCD camera in the AOI system to the standard mode: black level 110, gain 2.5 db, white balance 160 ± 5. Take the solar cell with normal film thickness (upper limit / lower limit / middle) for photographing and extract the HSV channel values and grayscale values of the card position images. Measure the first set of data: H channel value 155.8, S channel value 54.2, V channel value 36.9, gray grayscale value 20; Measure the second set of data: H channel value 185.8, S channel value 74.2, V channel value 56.9, gray grayscale value 20; Measure the third set of data: H channel value 162.4, S channel value 60.1, V channel value 43.1, gray grayscale value 20; In summary, take the average value combined with the deviation value to obtain the set ranges as follows: H channel value 170.8 ± 15, S channel value 64.2 ± 10, V channel value 46.9 ± 10, grayscale value 25 ± 5, number of card positions 3.

[0049] It can be understood that this method is based on the CCD camera to collect the solar cell images, and based on the electronic computer to judge the solar cell images, and output whether there is overplating on the solar cell. The output results can guide the automated production line to perform NG operations. It can reduce the error rate of online discrimination, enhance the recognition rate of the machine, and ensure the accuracy and efficiency of the judgment.

[0050] As shown in Table 1, the present application also provides specific experimental data to verify the recognition effect of this method. 110,000 battery wafers are taken from the product line and the traditional pure vision detection and the recognition method of this method are respectively performed, and the experimental data are as follows:

[0051] Table 1 Recognition data of battery wafers on the product line

[0052] Serial number Number of finished products Recognition quantity Accuracy rate False positive rate Finished product downgrade Not enabled 110000 117 64％ 4％ 0.03％ Enabled 110000 249 88％ 0％ 0.002％

[0053] It can be seen that after this method is enabled, the recognition accuracy rate is greatly improved and the misjudgment rate is significantly reduced. The present application has the advantages of accurate recognition area, high recognition efficiency and good recognition accuracy.

[0054] Please refer to Figure 4 shown Figure 4 which is a schematic structural diagram of a PECVD wrap plating recognition system based on the HSV model. The system includes: a data acquisition unit for acquiring the image of the battery wafer after coating collected by the CCD camera; a card position extraction unit for extracting all card position images from the battery wafer image; a determination data extraction unit for converting the card position image into the HSV color space to obtain the HSV channel values of the card position image, and graying the card position image to obtain the gray value of the card position image; a wrap plating determination unit for comparing the HSV channel values and gray values of the card position image with the set range to determine whether there is a card position, and comparing the number of card positions with the actual number of card positions of the battery wafer to determine whether there is wrap plating on the battery wafer.

[0055] In a possible implementation manner, the wrap plating determination unit includes: a set range extraction unit for acquiring the images of battery wafers with the film thickness at the normal upper limit, the film thickness at the normal lower limit, and the module at the normal median value collected by the standard mode CCD camera; extracting the card position images from the battery wafer images, and measuring the HSV channel values and gray values of the card position images; calculating the average value of the HSV channel values and the average value of the gray values of all the card position images as the reference values; setting a deviation value centered on the reference values to obtain the HSV channel set range and the gray value set range.

[0056] In a possible implementation manner, the wrap plating determination unit is specifically used for comparing the HSV channel values and gray values of the card position image with the set range. When both the HSV channel values and gray values of the card position image are within the set range, it is determined that there is a card position, marked as 1, otherwise marked as 0; counting the sum of the marks of all the card images on the same battery wafer image, and when the sum of the marks is equal to the actual number of card positions of the battery wafer, it is determined to be qualified, otherwise determined to be unqualified.

[0057] In a possible implementation manner, the card position extraction unit is specifically configured to extract the cell coordinate data from the cell image based on the AOI software; extract the rectangular frame coordinates centered on the card position based on the cell coordinate data; and extract the card position image from the cell image according to the rectangular frame coordinates.

[0058] In a possible implementation manner, the set range in the plating-around determination unit is: H channel value 170.8 ± 15, S channel value 64.2 ± 10, V channel value 46.9 ± 10, gray value 25 ± 5.

[0059] It can be understood that this system is used to execute the above method embodiments, corresponding one by one to the above method embodiments, and has the effects corresponding to the above method embodiments. Therefore, no further elaboration will be made here.

[0060] The above specific implementation manners have further elaborated in detail the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only the specific implementation manners of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A PECVD wrapping identification method based on the HSV model, characterized in that: include: Obtain the image of the cell after coating is completed captured by the CCD camera; Extract all card point images from the battery cell image; Converting the card point image into an HSV color space to obtain an HSV channel value of the card point image, and graying the card point image to obtain a gray value of the card point image; Compare the HSV channel value and grayscale value of the card point image with the set range to determine whether there is a card point, and compare the number of card points with the actual number of card points on the battery cell to determine whether there is bypass plating on the battery cell; The setting range is obtained by the following steps: obtaining a cell image with a film thickness within a normal upper limit, a film thickness within a normal lower limit, and a film thickness within a normal median value collected by a standard mode CCD camera; extracting a card point image from the cell image, and measuring the HSV channel value and grayscale value of the card point image; calculating the HSV channel value mean and grayscale value mean of all card point images as a reference value; setting a deviation value with the reference value as the center to obtain an HSV channel setting range and a grayscale value setting range.

2. The PECVD wrapping identification method based on the HSV model according to claim 1 is characterized in that: Comparing the HSV channel value and grayscale value of the card point image with the set range to determine whether there is a card point, and comparing the number of card points with the actual number of card points of the battery cell to determine whether the battery cell has bypass plating, including: Compare the HSV channel value and grayscale value of the card point image with the set range. When the HSV channel value and grayscale value of the card point image are both within the set range, it is determined that there is a card point and marked as 1, otherwise it is marked as 0; The identification sum of all the card point images on the same cell image is counted. When the identification sum is equal to the actual number of card point positions on the cell, it is judged as qualified, otherwise it is judged as unqualified.

3. The PECVD wrapping identification method based on the HSV model according to claim 1 is characterized in that: Extract all card point images from the battery cell image, including: Extracting cell coordinate data from the cell image based on AOI software; Extracting the coordinates of a rectangular frame centered at the card point based on the battery cell coordinate data; A card point image is extracted from the battery cell image according to the rectangular frame coordinates.

4. The PECVD wrapping identification method based on the HSV model according to claim 1 is characterized in that: The setting range is: H channel value 170.8±15, S channel value 64.2±10, V channel value 46.9±10, grayscale value 25±5.

5. A PECVD plating identification system based on the HSV model, characterized in that: include: A data acquisition unit, used to acquire the image of the cell after coating is completed collected by a CCD camera; A card point extraction unit, used to extract all card point images from the battery cell image; A determination data extraction unit is used to convert the card point image into an HSV color space to obtain an HSV channel value of the card point image, and grayscale the card point image to obtain a grayscale value of the card point image; A winding plating determination unit is used to compare the HSV channel value and grayscale value of the card point image with a set range to determine whether there is a card point, and compare the number of card points with the actual number of card points of the battery cell to determine whether the battery cell has winding plating; The winding plating determination unit includes: a setting range extraction unit for acquiring a cell image with a normal upper limit, a normal lower limit, and a normal median film thickness collected by a standard mode CCD camera; A card point image is extracted from the battery cell image, and the HSV channel value and grayscale value of the card point image are measured; the HSV channel value mean and the grayscale value mean of all card point images are calculated as reference values; a deviation value is set with the reference value as the center to obtain an HSV channel setting range and a grayscale value setting range.

6. A PECVD plating identification system based on the HSV model according to claim 5, characterized in that: The winding plating judgment unit is specifically used to compare the HSV channel value and grayscale value of the card point image with the set range. When the HSV channel value and grayscale value of the card point image are both within the set range, it is determined that there is a card point and marked as 1, otherwise it is marked as 0; the identification sum of all card point images on the same battery cell image is counted, and when the identification sum is equal to the actual number of card point positions on the battery cell, it is determined to be qualified, otherwise it is determined to be unqualified.

7. A PECVD wrapping identification system based on the HSV model according to claim 5, characterized in that: The card point extraction unit is specifically used to extract cell coordinate data from the cell image based on AOI software; extract rectangular frame coordinates centered on the card point based on the cell coordinate data; and extract a card point image from the cell image according to the rectangular frame coordinates.

8. The PECVD plating identification system based on the HSV model according to claim 5 is characterized in that: The setting range in the winding plating determination unit is: H channel value 170.8±15, S channel value 64.2±10, V channel value 46.9±10, and grayscale value 25±5.

Images