CN117788400A - Binocular vision-based photovoltaic module defect detection method and system - Google Patents
Binocular vision-based photovoltaic module defect detection method and system Download PDFInfo
- Publication number
- CN117788400A CN117788400A CN202311781749.XA CN202311781749A CN117788400A CN 117788400 A CN117788400 A CN 117788400A CN 202311781749 A CN202311781749 A CN 202311781749A CN 117788400 A CN117788400 A CN 117788400A
- Authority
- CN
- China
- Prior art keywords
- defect
- photovoltaic module
- dimensional
- detected
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000007547 defect Effects 0.000 title claims abstract description 157
- 238000001514 detection method Methods 0.000 title claims abstract description 34
- 238000004519 manufacturing process Methods 0.000 claims abstract description 47
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 41
- 238000013135 deep learning Methods 0.000 claims abstract description 34
- 238000000034 method Methods 0.000 claims abstract description 31
- 238000012545 processing Methods 0.000 claims abstract description 18
- 230000002950 deficient Effects 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 13
- 238000000605 extraction Methods 0.000 claims description 6
- 238000011426 transformation method Methods 0.000 claims description 6
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N Silicium dioxide Chemical group O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000003708 edge detection Methods 0.000 claims description 3
- 238000004590 computer program Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000013519 translation Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The invention relates to a binocular vision-based photovoltaic module defect detection method and system. The method comprises the steps of simultaneously acquiring left and right images of a photovoltaic module to be detected from different angles through two cameras; extracting characteristic points of the left image and the right image to obtain characteristic points of the left image and the right image and obtain parallax images of the two images; according to the parallax image, three-dimensional information of the surface of the photovoltaic module to be detected is obtained, and three-dimensional reconstruction is carried out on the photovoltaic module to be detected according to the three-dimensional information, so that a three-dimensional model is obtained; based on a three-dimensional model, acquiring quality and characteristic related parameters of the photovoltaic module to be detected; providing a deep learning algorithm model; and judging the defect grade and/or defect type of the photovoltaic module to be detected according to the deep learning algorithm model and combining the quality and characteristic related parameters, and generating a corresponding defect processing instruction to a production system according to a judging result. The invention can obviously improve the detection efficiency and accuracy of the defects of the photovoltaic module.
Description
Technical Field
The invention relates to the technical field of photovoltaic module manufacturing, in particular to a binocular vision-based photovoltaic module defect detection method and system.
Background
With the continuous transformation and upgrading of enterprise production technologies, automated technologies have played an increasingly important role in various links of the manufacturing industry. The automated production line constantly replaces the traditional manual operations during processing and assembly. However, in the production process of photovoltaic products, the generation mechanism of defects is relatively complex to the evaluation criteria. Currently, because of limitations in precision of conventional industrial cameras and machine vision software, manual visual inspection is relied upon in the industry to identify these defects.
Nevertheless, manual detection has significant drawbacks in terms of speed and recognition accuracy. Some defects are difficult to identify by the naked eye and human detectors often produce errors and missed decisions when subjectively judging the defect level. This not only affects the quality of the photovoltaic product, but also limits the production efficiency.
Disclosure of Invention
Therefore, the invention provides a binocular vision-based method and a binocular vision-based system for detecting defects of a photovoltaic module, which can remarkably improve the detection efficiency and the accuracy, can be rapidly adapted to different allowable standards by adjusting preset defect size limit values, and improves the detection efficiency and the quality of the photovoltaic module.
In order to solve the technical problems, the invention provides a binocular vision-based photovoltaic module defect detection method, which comprises the following steps:
simultaneously acquiring left and right images of the photovoltaic module to be detected from different angles through two cameras;
extracting characteristic points of the left image and the right image to obtain characteristic points of the left image and the right image and obtain parallax images of the two images;
according to the parallax image, three-dimensional information of the surface of the photovoltaic module to be detected is obtained, and three-dimensional reconstruction is carried out on the photovoltaic module to be detected according to the three-dimensional information, so that a three-dimensional model is obtained;
acquiring quality and characteristic related parameters of the photovoltaic module to be detected based on the three-dimensional model;
providing a deep learning algorithm model;
and judging the defect grade and/or defect type of the photovoltaic module to be detected according to the deep learning algorithm model and combining the quality and characteristic related parameters, and generating a corresponding defect processing instruction to a production system according to a judging result.
In one embodiment of the present invention, further comprising:
and acquiring production information of the photovoltaic module to be detected through a code reader while acquiring the left image and the right image, and storing the production information, the information obtained by combining the left image and the right image and the defect grade and/or defect type information.
In one embodiment of the present invention, the feature point extraction for the left and right images includes:
and extracting feature points of the left image and the right image by using a corner detection method, an edge detection method and a SURF feature extraction method.
In one embodiment of the present invention, obtaining three-dimensional information of the surface of the photovoltaic module to be detected according to the parallax map includes:
and calculating the three-dimensional information of the surface of the photovoltaic module to be detected by adopting a polar coordinate system transformation method or a cylindrical transformation method.
In one embodiment of the present invention, the quality and feature related parameters include: the photovoltaic module to be detected is characterized by comprising internal defects, appearance quality, geometric dimensions, surface flatness and angles of the module.
In one embodiment of the present invention, the deep learning algorithm model is obtained by:
manufacturing defect models with a plurality of limit sizes;
shooting a three-dimensional image of the defect model through an industrial camera;
measuring and adjusting errors of the actual size of the defect model and the corresponding three-dimensional image size, taking the actual size of the defect model as a defect sample, or recording the three-dimensional image size generated by a defect with a limit size as a defect sample;
selecting a deep learning algorithm, and inputting three-dimensional image data of the defect sample, wherein the three-dimensional image data comprise judging standards of no defects, different defect levels and/or defect types, so that the deep learning algorithm is learned and iterated to obtain a deep learning algorithm model;
and inputting brand new three-dimensional image data, and verifying the accuracy of the deep learning algorithm model in judging the defect grade and/or defect type.
In one embodiment of the present invention, the method for judging the defect grade and/or defect type of the photovoltaic module to be detected, and generating a corresponding defect processing instruction to a production system according to the judging result includes:
classifying said defect grade into defect free, light defect and severe defect;
if the photovoltaic module is judged to be defect-free, continuing to enter a subsequent production procedure;
if the photovoltaic module is judged to be slightly defective and the performance of the photovoltaic module is not affected by the slightly defective, a defective part photo is taken by identifying a bar code of the photovoltaic module, and a defect type label is added in a production system of the photovoltaic module so as to be traced back in a later period, and then the photovoltaic module enters a subsequent process;
if the photovoltaic module is judged to be a serious defect, the serious defect affects the performance and the use safety, a defective part photo is taken by identifying the bar code of the photovoltaic module, and a defect type label is added in a production system of the photovoltaic module so as to be convenient for later tracing, the photovoltaic module does not enter a subsequent process, and corresponding reworking or scrapping treatment is carried out according to the defect type.
In one embodiment of the present invention, the defect types include: the scratch of the battery piece is bad; breaking grid, scattering virtual printing and thick grid of battery grid line; poor sheet spacing, poor string length, misplacement of battery strings and misplacement of battery strings in the battery strings of the battery sheet arrangement; the frame is sunken; dirt and silica gel residue on the frame.
In one embodiment of the present invention, the deep learning algorithm employs a neural network algorithm.
The invention also provides a binocular vision-based defect detection system for the photovoltaic module, which comprises the following steps:
the image acquisition unit is used for simultaneously acquiring left and right images of the photovoltaic module to be detected from different angles through the two cameras;
the parallax image acquisition unit is used for extracting characteristic points of the left image and the right image to obtain characteristic points of the two images and obtain parallax images of the left image and the right image;
the three-dimensional model acquisition unit is used for obtaining three-dimensional information of the surface of the photovoltaic module to be detected according to the parallax map, and carrying out three-dimensional reconstruction on the photovoltaic module to be detected according to the three-dimensional information to obtain a three-dimensional model;
the quality and characteristic related parameter acquisition unit is used for acquiring quality and characteristic related parameters of the photovoltaic module to be detected based on the three-dimensional model;
a deep learning algorithm model unit for providing a deep learning algorithm model;
and the defect grading judgment unit is used for judging the defect grade and/or the defect type of the photovoltaic module to be detected according to the deep learning algorithm model by combining the quality and the characteristic related parameters, and generating a corresponding defect processing instruction to a production system according to a judgment result.
Compared with the prior art, the technical scheme of the invention has the following advantages:
according to the binocular vision-based photovoltaic module defect detection method and system, the efficiency and accuracy of distinguishing the photovoltaic defects can be improved, defect classification is automatically carried out, processing instructions are given, a machine and an algorithm are used for replacing manual inspection, two-dimensional picture information shot by a camera is synthesized into a three-dimensional image by means of binocular vision and three-dimensional reconstruction, the three-dimensional image information is imported into a trained algorithm, whether the three-dimensional image is defective or not is judged by the algorithm, the defect grade and type of the defect are judged by the algorithm, the corresponding processing instructions are output, different acceptance standards can be rapidly adapted, corresponding preset detection standard values are changed according to the different acceptance standards, and the rapid switching of the detection standards can be achieved.
Drawings
In order that the invention may be more readily understood, a more particular description of the invention will be rendered by reference to specific embodiments thereof that are illustrated in the appended drawings.
Fig. 1 is a flow chart of a method for detecting defects of a photovoltaic module based on binocular vision.
Fig. 2 is a layout diagram of the binocular vision-based defect detection system for the photovoltaic module.
Fig. 3 is a layout of the camera of the present invention.
Detailed Description
The present invention will be further described with reference to the accompanying drawings and specific examples, which are not intended to be limiting, so that those skilled in the art will better understand the invention and practice it.
In the present invention, if directions (up, down, left, right, front and rear) are described, they are merely for convenience of description of the technical solution of the present invention, and do not indicate or imply that the technical features must be in a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention.
In the present invention, "a plurality of" means one or more, and "a plurality of" means two or more, and "greater than", "less than", "exceeding", etc. are understood to not include the present number; "above", "below", "within" and the like are understood to include this number. In the description of the present invention, the description of "first" and "second" if any is used solely for the purpose of distinguishing between technical features and not necessarily for the purpose of indicating or implying a relative importance or implicitly indicating the number of technical features indicated or implicitly indicating the precedence of the technical features indicated.
In the present invention, unless clearly defined otherwise, terms such as "disposed," "mounted," "connected," and the like should be construed broadly and may be connected directly or indirectly through an intermediate medium, for example; the connecting device can be fixedly connected, detachably connected and integrally formed; can be mechanically connected, electrically connected or capable of communicating with each other; may be a communication between two elements or an interaction between two elements. The specific meaning of the words in the invention can be reasonably determined by a person skilled in the art in combination with the specific content of the technical solution.
Example 1
Referring to fig. 2 and 3, the photovoltaic module to be tested is conveyed through a production line (synchronous belt), a circular track is arranged on the periphery of the production line, the position of the photovoltaic module can be adjusted by moving the bottom guide rail along the direction of the production line, the left industrial camera and the right industrial camera are arranged on the circular track and positioned on the left side and the right side above the production line, a light source is arranged right above the circular track, and is used for collecting depth information of the photovoltaic module on the production line, and the position is unchanged during operation. And a code reader is also arranged on one side of the circular track, which is positioned in the flush direction of the assembly line.
Before the camera works, system debugging is needed, and the specific process is as follows: turning on the power supply of the camera and the detection system, and the light source. Entering a shooting mode, and observing whether dirt or other imaging defects exist on the image; adjusting the position of the camera to enable the area to be measured to be positioned at the center position in the view finding range of the camera; calibrating the camera to establish a consistent corresponding relation among an image coordinate system, a camera coordinate system and a world coordinate system; placing a grid plate in a region to be detected, ensuring that the region is completely covered, and then correcting the distortion of a camera; and (3) performing binocular stereo matching, determining the relative point-to-point relationship in the left image and the right image, further establishing parallax, and recovering the three-dimensional information of the points from the parallax.
During detection, the circular track is adjusted to a required detection position; before binocular vision detection, the two cameras need to be calibrated to determine internal parameters (e.g., focal length, optical center, distortion coefficients, etc.) and external parameters (e.g., rotation and translation matrices) of the cameras. These parameters may be obtained by manual calibration or automatic calibration methods. The two extrinsic matrices, the rotation matrix and the translation matrix, are generated during the calibration process, and the world coordinate system (the three-dimensional world coordinate system, which is introduced to describe the position of the object in the real world) is transformed by the translation matrix and the rotation matrix and then with the camera coordinate system. The calibration aims at determining the conversion relation (internal and external parameters) between the three-dimensional space point and the pixel point of the pixel plane under the world coordinate system and determining the distortion system in the imaging process of the camera so as to correct the image.
Referring to fig. 1, a method for detecting defects of a photovoltaic module based on binocular vision includes:
s1, simultaneously acquiring left and right images of a photovoltaic module to be detected from different angles through two cameras, wherein the images comprise information such as the surface, the edge and the structure of the photovoltaic module, acquiring production information of the photovoltaic module to be detected through a code reader while acquiring the left and right images, combining the production information with the left and right images, and storing the production information.
The production information of the photovoltaic module being inspected is associated with the acquired image and stored, and the two-dimensional code information on the photovoltaic module is read and combined with the corresponding image, so that the traceability and the recording of each module can be realized, the production, the quality control and the subsequent traceability of the photovoltaic module can be tracked, and the quality and the traceability of the photovoltaic module can be ensured.
S2, extracting characteristic points of the left image and the right image to obtain characteristic points of the left image and the right image and obtain parallax images of the two images; the feature point extraction method can adopt a corner detection method, an edge detection method and a SURF feature extraction method to extract feature points; the parallax calculation can be realized by adopting a polar coordinate system transformation method or a cylindrical transformation method and the like.
And S3, obtaining three-dimensional information of the surface of the photovoltaic module to be detected according to the parallax map, and carrying out three-dimensional reconstruction on the photovoltaic module to be detected according to the three-dimensional information to obtain a three-dimensional model.
And S4, based on the three-dimensional model, acquiring quality and characteristic related parameters of the photovoltaic module to be detected through analysis and processing of the three-dimensional model, wherein the parameters comprise internal defects, appearance quality, geometric dimensions, surface flatness and angles of the module of the photovoltaic module to be detected.
Step S5, providing a deep learning algorithm model, wherein the deep learning algorithm model is obtained by the following steps:
manufacturing defect models with a plurality of limit sizes; the limit size is the maximum value and the minimum value of the defect related parameters of the defect model;
shooting a three-dimensional image of the defect model through an industrial camera;
measuring and adjusting errors of the actual size of the defect model and the corresponding three-dimensional image size, taking the actual size of the defect model as a defect sample, or recording the three-dimensional image size generated by a defect with a limit size as a defect sample;
selecting a deep learning algorithm, such as a neural network algorithm, inputting three-dimensional image data of the defect sample, including judging standards of no defect, different defect levels and/or defect types, and enabling the deep learning algorithm to learn and iterate to obtain a deep learning algorithm model;
and inputting brand new three-dimensional image data, and verifying the accuracy of the deep learning algorithm model in judging the defect grade and/or defect type. The accuracy rate checking process can be regularly carried out after a period of actual use so as to ensure the detection quality.
And S6, judging the defect grade and/or defect type of the photovoltaic module to be detected according to the deep learning algorithm model by combining the quality and the characteristic related parameters, and generating a corresponding defect processing instruction to a production system according to a judging result.
In this embodiment, the bar code of the defective component, the defect level, the defect type information, and the defect instruction are uploaded to the production system, and all the above process information is stored by the information storage module.
The present example will classify the defect class into defect free, slight defect and severe defect; the defect processing instruction includes:
follow current: continuing to enter the subsequent production process if the defect is not detected;
grading: slight defects do not affect the performance of the photovoltaic module, the bar codes of the photovoltaic module are identified, the defective parts are photographed, defect type labels are added in a production system, the later tracing is facilitated, and the follow-up production process is carried out;
reworking and scrapping: serious defects influence the performance and the use safety of the photovoltaic module, identify the bar code of the photovoltaic module and photograph the defective part, and add defect type labels in a production system so as to facilitate the later tracing; and carrying out reworking or scrapping flow according to the corresponding defect type and defect degree and a preset scheme.
The process comprises the following steps: if the photovoltaic module is judged to be defect-free, continuing to enter a subsequent production procedure; if the photovoltaic module is judged to be slightly defective and the performance of the photovoltaic module is not affected by the slightly defective, a defective part photo is taken by identifying a bar code of the photovoltaic module, and a defect type label is added in a production system so as to be traced back in the later period, and then the photovoltaic module enters a subsequent process; if the photovoltaic module is judged to be a serious defect, the serious defect affects the performance and the use safety, a defect part photo is taken by identifying the bar code of the photovoltaic module, and a defect type label is added in a production system so as to trace back later, the photovoltaic module does not enter a subsequent process, and corresponding reworking or scrapping treatment is carried out according to the defect type.
In this embodiment, the defect types include: the scratch of the battery piece is bad; breaking grid, scattering virtual printing and thick grid of battery grid line; poor sheet spacing, poor string length, misplacement of battery strings and misplacement of battery strings in the battery strings of the battery sheet arrangement; the frame is sunken and the frame is installed in a staggered manner; dirt and silica gel residue on the frame. The corresponding defect handling instructions are shown in table 1.
TABLE 1
Example 2
Based on the same inventive concept, the present embodiment provides a binocular vision-based defect detection system for a photovoltaic module, and the principle of solving the problem is similar to that of the binocular vision-based defect detection method, and the repetition is not repeated.
The embodiment provides a photovoltaic module defect detection system based on binocular vision, including:
the image acquisition unit is used for simultaneously acquiring left and right images of the photovoltaic module to be detected from different angles through the two cameras;
the parallax image acquisition unit is used for extracting characteristic points of the left image and the right image to obtain characteristic points of the two images and obtain parallax images of the left image and the right image;
the three-dimensional model acquisition unit is used for obtaining three-dimensional information of the surface of the photovoltaic module to be detected according to the parallax map, and carrying out three-dimensional reconstruction on the photovoltaic module to be detected according to the three-dimensional information to obtain a three-dimensional model;
the quality and characteristic related parameter acquisition unit is used for acquiring quality and characteristic related parameters of the photovoltaic module to be detected based on the three-dimensional model;
a deep learning algorithm model unit for providing a deep learning algorithm model;
and the defect grading judgment unit is used for judging the defect grade and/or the defect type of the photovoltaic module to be detected according to the deep learning algorithm model by combining the quality and the characteristic related parameters, and generating a corresponding defect processing instruction to a production system according to a judgment result.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that the above-mentioned embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same, and although the present invention has been described in detail with reference to examples, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention, and all such modifications and equivalents are intended to be encompassed in the scope of the claims of the present invention.
Claims (10)
1. The method for detecting the defects of the photovoltaic module based on binocular vision is characterized by comprising the following steps of:
simultaneously acquiring left and right images of the photovoltaic module to be detected from different angles through two cameras;
extracting characteristic points of the left image and the right image to obtain characteristic points of the left image and the right image and obtain parallax images of the two images;
according to the parallax image, three-dimensional information of the surface of the photovoltaic module to be detected is obtained, and three-dimensional reconstruction is carried out on the photovoltaic module to be detected according to the three-dimensional information, so that a three-dimensional model is obtained;
acquiring quality and characteristic related parameters of the photovoltaic module to be detected based on the three-dimensional model;
providing a deep learning algorithm model;
and judging the defect grade and/or defect type of the photovoltaic module to be detected according to the deep learning algorithm model and combining the quality and characteristic related parameters, and generating a corresponding defect processing instruction to a production system according to a judging result.
2. The binocular vision-based photovoltaic module defect detection method of claim 1, further comprising:
and acquiring production information of the photovoltaic module to be detected through a code reader while acquiring the left image and the right image, and storing the production information, the information obtained by combining the left image and the right image and the defect grade and/or defect type information.
3. The binocular vision-based photovoltaic module defect detection method of claim 1, wherein the feature point extraction is performed on the left and right images, and the method comprises the following steps:
and extracting feature points of the left image and the right image by using a corner detection method, an edge detection method and a SURF feature extraction method.
4. The binocular vision-based photovoltaic module defect detection method of claim 1, wherein obtaining the three-dimensional information of the surface of the photovoltaic module to be detected according to the parallax map comprises:
and calculating the three-dimensional information of the surface of the photovoltaic module to be detected by adopting a polar coordinate system transformation method or a cylindrical transformation method.
5. The binocular vision-based photovoltaic module defect detection method of claim 1, wherein the quality and feature related parameters comprise: the photovoltaic module to be detected is characterized by comprising internal defects, appearance quality, geometric dimensions, surface flatness and angles of the module.
6. The binocular vision-based photovoltaic module defect detection method of claim 1, wherein the deep learning algorithm model is obtained by the following method:
manufacturing defect models with a plurality of limit sizes;
shooting a three-dimensional image of the defect model through an industrial camera;
measuring and adjusting errors of the actual size of the defect model and the corresponding three-dimensional image size, taking the actual size of the defect model as a defect sample, or recording the three-dimensional image size generated by a defect with a limit size as a defect sample;
selecting a deep learning algorithm, and inputting three-dimensional image data of the defect sample, wherein the three-dimensional image data comprise judging standards of no defects, different defect levels and/or defect types, so that the deep learning algorithm is learned and iterated to obtain a deep learning algorithm model;
and inputting brand new three-dimensional image data, and verifying the accuracy of the deep learning algorithm model in judging the defect grade and/or defect type.
7. The binocular vision-based photovoltaic module defect detection method according to claim 1, wherein the method for judging the defect grade and/or the defect type of the photovoltaic module to be detected and generating a corresponding defect processing instruction to a production system according to the judging result comprises the following steps:
classifying said defect grade into defect free, light defect and severe defect;
if the photovoltaic module is judged to be defect-free, continuing to enter a subsequent production procedure;
if the photovoltaic module is judged to be slightly defective and the performance of the photovoltaic module is not affected by the slightly defective, a defective part photo is taken by identifying a bar code of the photovoltaic module, and a defect type label is added in a production system so as to be traced back in the later period, and then the photovoltaic module enters a subsequent process;
if the photovoltaic module is judged to be a serious defect, the serious defect affects the performance and the use safety, a defect part photo is taken by identifying the bar code of the photovoltaic module, and a defect type label is added in a production system so as to trace back later, the photovoltaic module does not enter a subsequent process, and corresponding reworking or scrapping treatment is carried out according to the defect type.
8. The binocular vision-based photovoltaic module defect detection method of claim 1, wherein the defect types include: the scratch of the battery piece is bad; breaking grid, scattering virtual printing and thick grid of battery grid line; poor sheet spacing, poor string length, misplacement of battery strings and misplacement of battery strings in the battery strings of the battery sheet arrangement; the frame is sunken and the frame is installed in a staggered manner; dirt and silica gel residue on the frame.
9. The binocular vision-based photovoltaic module defect detection method of claim 6, wherein the deep learning algorithm adopts a neural network algorithm.
10. A binocular vision-based photovoltaic defect detection system, comprising:
the image acquisition unit is used for simultaneously acquiring left and right images of the photovoltaic module to be detected from different angles through the two cameras;
the parallax image acquisition unit is used for extracting characteristic points of the left image and the right image to obtain characteristic points of the two images and obtain parallax images of the left image and the right image;
the three-dimensional model acquisition unit is used for obtaining three-dimensional information of the surface of the photovoltaic module to be detected according to the parallax map, and carrying out three-dimensional reconstruction on the photovoltaic module to be detected according to the three-dimensional information to obtain a three-dimensional model;
the quality and characteristic related parameter acquisition unit is used for acquiring quality and characteristic related parameters of the photovoltaic module to be detected based on the three-dimensional model;
a deep learning algorithm model unit for providing a deep learning algorithm model;
and the defect grading judgment unit is used for judging the defect grade and/or the defect type of the photovoltaic module to be detected according to the deep learning algorithm model by combining the quality and the characteristic related parameters, and generating a corresponding defect processing instruction to a production system according to a judgment result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311781749.XA CN117788400A (en) | 2023-12-22 | 2023-12-22 | Binocular vision-based photovoltaic module defect detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311781749.XA CN117788400A (en) | 2023-12-22 | 2023-12-22 | Binocular vision-based photovoltaic module defect detection method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117788400A true CN117788400A (en) | 2024-03-29 |
Family
ID=90397487
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311781749.XA Pending CN117788400A (en) | 2023-12-22 | 2023-12-22 | Binocular vision-based photovoltaic module defect detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117788400A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118347942A (en) * | 2024-06-17 | 2024-07-16 | 华羿微电子股份有限公司 | Method, equipment and storage medium for detecting appearance of semiconductor product after rib cutting molding |
-
2023
- 2023-12-22 CN CN202311781749.XA patent/CN117788400A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118347942A (en) * | 2024-06-17 | 2024-07-16 | 华羿微电子股份有限公司 | Method, equipment and storage medium for detecting appearance of semiconductor product after rib cutting molding |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110659660B (en) | Automatic optical detection classification equipment using deep learning system and training equipment thereof | |
| CN111091538B (en) | Automatic identification and defect detection method and device for pipeline welding seams | |
| US11315229B2 (en) | Method for training defect detector | |
| CN111982921A (en) | Hole defect detection method and device, conveying platform and storage medium | |
| CN112884743B (en) | Detection method and device, detection equipment and storage medium | |
| CN113325001A (en) | Automatic distinguishing and detecting equipment and method for surface appearance flaws of copper foil | |
| CN114529510B (en) | Automatic detection and classification method for cathode copper on-line quality | |
| CN113222913B (en) | Circuit board defect detection positioning method, device and storage medium | |
| CN113935971A (en) | Method and device for detecting surface defects of composite material | |
| CN109816634A (en) | Detection method, model training method, device and equipment | |
| CN117788400A (en) | Binocular vision-based photovoltaic module defect detection method and system | |
| CN111929239A (en) | AOI detection device and detection method for PCB part defects | |
| CN114494103A (en) | Defect detection method and device | |
| CN116091506B (en) | Machine vision defect quality inspection method based on YOLOV5 | |
| CN105572133A (en) | Flaw detection method and device | |
| CN112763496A (en) | Mobile phone battery surface defect detection device and detection method thereof | |
| CN117635599A (en) | Defect detection model training and defect detection method, device, equipment and medium | |
| CN116681677A (en) | Lithium battery defect detection method, device and system | |
| CN116071365A (en) | Part detection method, device, equipment and storage medium | |
| CN117036231A (en) | Visual detection method for spot welding quality of automobile cabin assembly | |
| CN112884691B (en) | Data enhancement device, data enhancement apparatus, and storage medium | |
| CN115937555A (en) | Industrial defect detection algorithm based on standardized flow model | |
| CN107123105A (en) | Images match defect inspection method based on FAST algorithms | |
| CN113267506A (en) | Wood board AI visual defect detection device, method, equipment and medium | |
| CN114577816A (en) | Hydrogen fuel bipolar plate detection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |