EVA small strip visual calibration method and device

Abstract

The invention discloses an EVA small strip visual calibration method and device, and relates to the technical field of visual calibration. According to the method, two vision cameras are adopted to shoot a calibration plate, and a fixed coordinate system is fitted; eVA small strip photos are shot at different rotation angles, and the rotation center and the rotation position of the manipulator in the fitting coordinate system are calculated; carrying out picture frame processing on the pictures shot by the two visual cameras, and identifying and acquiring an image of the EVA small strip to be detected; the image is used as a template for subsequent deviation correction processing. According to the EVA small strip visual calibration method and device provided by the invention, the detection mode is converted from traditional edge detection to minimum rectangle identification, the crossing from line finding to surface finding is realized, the false alarm rate is significantly reduced, the detection precision is improved, and object identification and measurement are more accurate and efficient.

Description

Technical Field

[0001] The present application belongs to the field of visual calibration technology, and in particular relates to a method and device for visual calibration of EVA strips. Background Art

[0002] In the photovoltaic module industry, EVA (Ethylene-Vinyl Acetate) strips are a key encapsulation material, primarily used for encapsulating solar cells to ensure the stability and durability of photovoltaic modules. Calipers are often used to calibrate objects and identify edges.

[0003] However, this approach has the following problems:

[0004] First, when there is debris around the material, the caliper may fail to find the edge due to the bevel phenomenon, affecting the accuracy of the measurement; secondly, if the material is placed beyond the detection range of the caliper, it may also lead to recognition failure and ineffective positioning and measurement. Summary of the Invention

[0005] To address the problems in the prior art, this application provides a method and device for visual calibration of EVA strips, which not only significantly reduces the false alarm rate but also improves detection accuracy, making object recognition and measurement more accurate and efficient. This ensures stable and reliable detection even in complex scenarios.

[0006] The technical solution is as follows:

[0007] In one aspect, a method for visually calibrating an EVA strip is provided, comprising:

[0008] Step S1: Use two visual cameras to shoot the calibration plate and fit a fixed coordinate system;

[0009] Step S2: Take photos of the EVA strip at different rotation angles and calculate the rotation center and rotation position of the manipulator in the fitting coordinate system;

[0010] Step S3: Frame the photos taken by the two visual cameras to identify and obtain the image of the EVA strip to be detected; the image will be used as a template for subsequent correction processing.

[0011] A further technical solution is to draw the frame in the following steps:

[0012] Binarize the image;

[0013] Apply Gaussian filtering to remove noise from the image;

[0014] Use a mask to block out distracting parts in the middle of the image;

[0015] Use morphological operations to remove the remaining noise content in the frame part;

[0016] Snap the smallest rectangular outline of the EVA strip.

[0017] As a further technical solution, in step S3, the area of the EVA strip to be detected in the image is obtained through an attention frame selection mechanism.

[0018] A further technical solution is to select EVA strips at different positions in the image based on the attention selection mechanism, and compensate for the parameters of EVA strips at different positions.

[0019] According to a further technical solution, step S3 further includes calculating an average value of the offsets of the X-axis, Y-axis, and U mechanical axis based on the offsets of the X-axis, Y-axis, and U mechanical axis to achieve deviation correction.

[0020] A further technical solution is to calculate the rotation center of the manipulator through an algorithm based on taking photos of the manipulator at different rotation angles and three offset positions of the three rotation angles.

[0021] A further technical solution is to take a photo at three different angles, and use an algorithm to calculate the intersection point of the upper left corner of the rectangle at each angle, obtaining three intersection points of the upper left corner and the corresponding three intersection points of the upper right corner. Then, the rotation center is fitted based on the upper left and upper right intersection points of the EVA strips in the pictures at different angles. The specific steps are:

[0022] Assume there are two sets of EVA strips in the picture, each picture has two intersection points, each angle has two left and right pictures, a total of three angles, a total of 18 points; these 18 points are divided into three groups, each group is the position of the corresponding point at different rotation angles;

[0023] Assume that the first set of points is P1=(x1,y1),(x2,y2),…,(x6,y 16 ), the second set of points is P2=(x′1,y′1),(x′2,y′2),…,(x′6,y′6), where each pair (x i ,y i ) and (x′ i ,y′ i )correspond;

[0024] Assume that the rotation is a rotation in two-dimensional space, and the center of rotation is (x c ,y c ), the rotation angle is θ, and the rotation formula is expressed as:

[0025]

[0026] Among them, (x i ,yi ) is the first set of points, (x′ i ,y′ i ) is the second set of points after rotation;

[0027] By expanding the above rotation equation, for each point i, we get the following equation:

[0028] x′ i -x c =cos(θ)×(x i -x c )-sin(θ)×(y i -y c )

[0029] y′ i -y c =sin(θ)×(x i -x c )+cos(θ)×(y i -y c )

[0030] Build a system of equations through all the points and solve for the center of rotation (x c ,y c ) and the rotation angle θ;

[0031] The rotation angle and rotation center are fitted by the least square method to solve the rotation angle θ and the rotation center (x c ,y c ).

[0032] According to a further technical solution, the method is applicable to detecting any type of film including a film having a length of one strip or more, a film with holes, a film with white stripes and a black background, and a film with black stripes and a white background.

[0033] On the other hand, a device for visually calibrating an EVA strip is provided, which adopts the above-mentioned method for visually calibrating an EVA strip, and the device comprises:

[0034] The calibration module is used to use two visual cameras to shoot the calibration plate and fit a fixed coordinate system;

[0035] Calibration module: used to take photos of the EVA strip at different rotation angles and calculate the rotation center and rotation position of the manipulator in the fitting coordinate system;

[0036] Determination module: used to perform frame processing and recognition on the photos taken by the two visual cameras and obtain the image of the EVA strip to be detected; the image will be used as a template for subsequent correction processing.

[0037] The technical solution includes at least the following technical effects:

[0038] This application shifts detection from traditional edge detection to minimum rectangle recognition, achieving a leap from "line finding" to "surface finding." This significantly reduces false positives and improves detection accuracy, making object recognition and measurement more precise and efficient. This technical solution enables precise positioning and measurement even when materials are placed beyond the detection range of traditional measuring tools.

[0039] At the same time, an attention frame selection mechanism is employed to effectively filter out external interference factors such as debris, significantly improving the stability and accuracy of detection. This attention mechanism significantly expands the recognition range, no longer limited to the caliper's detection range, to cover the entire image range, thereby improving detection precision and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0041] Figure 1 A flow chart of a visual calibration method for EVA strips provided in a preferred embodiment of the present application;

[0042] Figure 2 A schematic structural diagram of an EVA strip visual calibration device provided in a preferred embodiment of the present application;

[0043] Figure 3 This is the picture frame area in the calibration plate provided in a preferred embodiment of the present application. DETAILED DESCRIPTION

[0044] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0045] Example 1

[0046] Attachment Figure 1 A flow chart of a visual calibration method for EVA strips provided in a preferred embodiment is shown in the figure.

[0047] Step S1: Use two visual cameras to shoot the calibration plate, and fit the fixed points on the calibration plate to a fixed coordinate.

[0048] In this embodiment, a 150cm*15cm checkerboard (calibration board) is placed on the photographing platform. The left and right visual cameras each take a picture to obtain two pictures. The fixed points on the calibration board are fitted to a fixed coordinate. The specific steps are as follows:

[0049] Take a picture with the left visual camera, and use the algorithm to realize the frame step. Select a rectangular frame from the upper and lower frames. There are 9 intersection points between black and white in each rectangular frame (such as Figure 3 As shown in the figure, there are 18 intersection points in the upper and lower groups. Similarly, the right visual camera takes a photo and performs the framing step, also obtaining 18 intersection points, for a total of 36 points. The left and right coordinates are integrated into the same coordinate system, and the coordinate systems of the two visual cameras are integrated into the same coordinate system to establish a unified reference plane.

[0050] Step S2: Take photos at different rotation angles, calculate the rotation center through the algorithm, and obtain the rotation position of the manipulator in the fitted coordinate system;

[0051] Positive and negative rotation of a certain angle, three offset positions of different angles. In this embodiment, the left and right visual cameras are fixed on the left and right sides of the tooling platform respectively, and the calibration plate is placed on the tooling platform. The robot is translated by the servo motor. In step S1, the calibration plate is placed on the tooling platform. In step S2, the calibration plate is removed and the EVA strip is placed on the tooling platform. A photo is taken at three different angles: 0°, -1°, and +1°. The algorithm is used to calculate the intersection point of the upper left corner of the rectangle at each angle, resulting in three upper left corner intersection points and the corresponding three upper right corner intersection points. Then, the rotation center is fitted based on the upper left and upper right intersection points of the EVA strip in the pictures at different angles. The specific steps are as follows:

[0052] Suppose there are two sets of EVA strips in the image, each with two intersection points, and each angle has two left and right images, for a total of three angles, for a total of 18 points. These 18 points are divided into three groups, each group is the position of the corresponding point at different rotation angles.

[0053] Assume that the first set of points is P1=(x1,y1),(x2,y2),…,(x6,y6), and the second set of points is P2=(x′1,y′1),(x′2,y′2),…,(x′6,y′6), where each pair (x i ,y i ) and (x′ i ,y′ i )correspond;

[0054] Assume that the rotation is a rotation in two-dimensional space, and the center of rotation is (x c ,y c ), the rotation angle is θ, and the rotation formula is expressed as:

[0055]

[0056] Among them, (x i ,y i ) is the first set of points, (x′ i ,y′ i ) is the second set of points after rotation;

[0057] By expanding the above rotation equation, for each point i, we get the following equation:

[0058] x′ i -x c =cos(θ)×(x i -x c )-sin(θ)×(y i -y c )

[0059] y′ i -y c =sin(θ)×(x i -x c )+cos(θ)×(y i -y c )

[0060] Build a system of equations through all the points and solve for the center of rotation (x c ,y c ) and the rotation angle θ;

[0061] The rotation angle and rotation center are fitted by the least square method to solve the rotation angle θ and the rotation center (x c ,y c ).

[0062] Step S3: The images captured by the two visual cameras are framed to identify the image and position of the EVA strip to be inspected. This image serves as a template for subsequent deflection correction, and compensation is performed on the first piece. Based on the offsets of the X, Y, and U mechanical axes, the average of these offsets is calculated to achieve deflection correction. Communication services are used for data transmission.

[0063] The steps of the frame algorithm are as follows: for the frame part, first, the image is binarized to highlight the target outline; then, Gaussian filtering technology is applied to eliminate the noise in the image and improve the image quality; then, a masking frame is used to shield the interference part in the middle of the image to ensure the accuracy of the algorithm; then, morphological operations are used to remove the remaining noise content in the frame part (including opening, closing, erosion, etc.), and finally, the minimum rectangular outline of the EVA strip is captured.

[0064] The attention box selection mechanism captures the region of the EVA strip to be detected in the image and obtains the photo RoI. Left and right offsets are given, and the average of the X, Y, and U offsets is taken. This average is used to account for any bulges in the center of the strip.

[0065] Based on the attention selection mechanism, EVA strips at different positions in the image are selected, and the parameters of the EVA strips at different positions are compensated, and the original template parameters are adjusted to meet actual needs. In this embodiment, the head, middle, and tail of the EVA strip are selected and the parameters of the head, middle, and tail of the EVA strip are compensated.

[0066] The EVA strip visual calibration method of this embodiment is suitable for detecting any type of film including a film that is longer than or equal to a strip, a film with holes, a film with a white stripe and a black background, and a film with a black stripe and a white background (black silicon component).

[0067] In one embodiment, a circular hole is taken as an example and used as a reference.

[0068] Example 2

[0069] Attachment Figure 2 A schematic diagram of the structure of an EVA strip visual calibration device is provided as a preferred embodiment. Figure 2 As shown, a visual calibration device for EVA strips adopts the above-mentioned visual calibration method for EVA strips, and the device includes:

[0070] The calibration module is used to use two visual cameras to shoot the calibration plate and fit a fixed coordinate system;

[0071] Calibration module: used to take photos of the EVA strip at different rotation angles and calculate the rotation center and rotation position of the manipulator in the fitting coordinate system;

[0072] Determination module: used to perform frame processing on the photos taken by the two visual cameras, identify and obtain the image of the EVA strip to be detected; the image will be used as a template for subsequent correction processing.

[0073] The EVA strip visual calibration method of this application has the following technical effects:

[0074] 1. Improve detection accuracy: By using two visual cameras to shoot the calibration plate and fit a fixed coordinate system, the accuracy and stability of the calibration are ensured, providing a precise reference system for subsequent detection.

[0075] 2. Precise rotation positioning: By taking photos of the EVA strip at different rotation angles, the rotation center and rotation position of the manipulator in the fitting coordinate system are calculated, achieving precise control and positioning of the manipulator's rotation movement.

[0076] 3. Effective noise removal: The frame algorithm effectively eliminates noise and interference in the image through binarization, Gaussian filtering, masking, and morphological operations, ensuring that the minimum rectangular outline of the EVA strip can be accurately captured.

[0077] 4. Enhanced anti-interference capability: The introduction of an attention frame selection mechanism can effectively avoid interference from external factors such as debris, significantly improving the stability and accuracy of detection.

[0078] 5. Expand the detection range: With the help of the attention mechanism, the detection range is no longer limited to the detection range of traditional tools, but can cover the entire image range, thereby obtaining more comprehensive and accurate detection results.

[0079] 6. Achieve precise deviation correction: By calculating the average value of the offset of the X-axis, Y-axis, and U-axis, precise deviation correction of the EVA strip is achieved, ensuring accurate alignment and processing of materials during the production process.

[0080] 7. Adaptability to complex scenarios: Even when the material placement is beyond the detection range of traditional measuring tools, this method can achieve accurate positioning and measurement, improving the flexibility and adaptability of the production process.

[0081] To sum up, the EVA strip visual calibration method and device provided in this application have significant technical effects in improving detection accuracy, enhancing anti-interference ability, expanding detection range and achieving precise correction, and can effectively improve production efficiency and product quality.

[0082] This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the present application being indicated by the appended claims.

[0083] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A visual calibration method for EVA strips, characterized in that: include: Step S1: Use two visual cameras to shoot the calibration plate and fit a fixed coordinate system; Step S2: Take photos of the EVA strip at different rotation angles and calculate the rotation center and rotation position of the manipulator in the fitting coordinate system; Step S3: Frame the photos taken by the two visual cameras to identify and obtain the image of the EVA strip to be detected; the image will be used as a template for subsequent correction processing.

2. The EVA strip visual calibration method according to claim 1, characterized in that: The steps of the picture frame are: Binarize the image; Apply Gaussian filtering to remove noise from the image; Use a mask to block out distracting parts in the middle of the image; Use morphological operations to remove the remaining noise content in the frame part; Snap the smallest rectangular outline of the EVA strip.

3. The EVA strip visual calibration method according to claim 1, characterized in that: In step S3, the area of the EVA strip to be detected in the image is obtained through the attention frame selection mechanism.

4. The EVA strip visual calibration method according to claim 3, characterized in that: Based on the attention selection mechanism, EVA strips at different positions in the image are selected and the parameters of EVA strips at different positions are compensated.

5. The EVA strip visual calibration method according to claim 4, characterized in that: The step S3 further includes calculating an average value of the offsets of the X-axis, Y-axis, and U-mechanical axis based on the offsets of the X-axis, Y-axis, and U-mechanical axis to achieve deviation correction.

6. The EVA strip visual calibration method according to claim 1, characterized in that: The method is based on taking photos of the manipulator at different rotation angles, rotating three offset positions at three angles, and calculating the rotation center of the manipulator through an algorithm.

7. The EVA strip visual calibration method according to claim 1, characterized in that: Take a photo at three different angles and use the algorithm to calculate the intersection point of the upper left corner of the rectangle at each angle. A total of three intersection points of the upper left corner and the corresponding three intersection points of the upper right corner are obtained. Then, the rotation center is fitted based on the upper left and upper right intersection points of the EVA strip in the pictures at different angles. The specific steps are: Assume there are two sets of EVA strips in the picture, each picture has two intersection points, each angle has two left and right pictures, a total of three angles, a total of 18 points; these 18 points are divided into three groups, each group is the position of the corresponding point at different rotation angles; Assume that the first set of points is P1=(x1,y1),(x2,y2),…,(x6,y6), and the second set of points is P2=(x′1,y′1),(x′2,y′2),…,(x′6,y′6), where each pair (x i ,y i ) and (x′ i ,y′ i )correspond; Assume that the rotation is a rotation in two-dimensional space, and the center of rotation is (x c ,y c ), the rotation angle is θ, and the rotation formula is expressed as: Among them, (x i ,y i ) is the first set of points, (x′ i ,y′ i ) is the second set of points after rotation; By expanding the above rotation equation, for each point i, we get the following equation: x′ i -x c =cos(θ)×(x i -x c )-sin(θ)×(y i -and c ) and' i -and c =sin(θ)×(x i -x c )+cos(θ)×(y i -and c ) Build a system of equations through all the points and solve for the center of rotation (x c ,y c ) and the rotation angle θ; The rotation angle and rotation center are fitted by the least square method to solve the rotation angle θ and the rotation center (x c ,y c ).

8. The EVA strip visual calibration method according to claim 1, characterized in that: The method is applicable to detecting any type of adhesive film including an adhesive film having a length of one strip or more, an adhesive film with holes, an adhesive film with white stripes and a black background, and an adhesive film with black stripes and a white background.

9. An EVA strip visual calibration device, using the EVA strip visual calibration method according to any one of claims 1 to 8, characterized in that: The device comprises: The calibration module is used to use two visual cameras to respectively photograph the calibration plate and fit the fixed points on the calibration plate to a fixed coordinate system; Calibration module: used to take photos of the EVA strip at different rotation angles, calculate the rotation center through the algorithm, and obtain the rotation center of the manipulator in the fitted coordinate system; Determination module: used to perform frame processing on the photos taken by the two visual cameras through the frame algorithm to identify and obtain the image of the EVA strip to be detected; the image will be used as a template for subsequent correction processing.

Images