Photovoltaic defect remote monitoring method based on wireless communication transmission
Through wireless communication transmission technology, the first and second follow-up monitoring components are used to generate and identify characteristic images of photovoltaic modules, and the problem of large computing power occupation and low detection probability in small-scale defect detection of photovoltaic modules is solved, thus achieving high efficiency of photovoltaic defects Remote monitoring and identification.
Patent Information
- Application Number
- CN202411320392.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-23
- Publication Date
- 2024-12-20
- Estimated Expiration
- 2044-09-23
AI Technical Summary
In the prior art, small-scale defect detection of photovoltaic modules requires a large amount of computing power and storage resources, and the detection probability is low, resulting in redundant computing power and storage occupation problems.
Using a remote monitoring method for photovoltaic defects based on wireless communication transmission, the grayscale characteristic image of the photovoltaic component is generated by the first follow-up monitoring component and transmitted to the cloud, image segmentation is performed to obtain a quasi-defect identification area, and then aligned using the second follow-up monitoring component. The defect area is followed and photographed at the same speed, generate rich feature images and identify defects to achieve defect detection.
It improves the probability of photovoltaic module defect detection, reduces computing power redundancy, and realizes efficient detection of remote monitoring of photovoltaic defects.
Smart Images
Figure CN118868787B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of wireless communications, and in particular to a photovoltaic defect remote monitoring method based on wireless communication transmission. Background Art
[0002] The current detection method for some small-scale defects of photovoltaic modules, such as cracks, hidden cracks, broken grids, and black spots, is achieved by combining high-speed cameras with target detection algorithms. The disadvantage is that both the process of high-speed camera image acquisition and the process of target detection algorithms for small-scale defect detection require a large number of data nodes, resulting in a huge amount of memory and operating space. However, in actual quality inspection, the probability of defect detection may be relatively low, resulting in more redundant computing power and storage usage. How to reduce redundant computing power and storage usage while ensuring defect detection has become a major problem in current industrial applications. Summary of the invention
[0003] The present application provides a photovoltaic defect remote monitoring method based on wireless communication transmission, which is used to solve the technical problems of small-scale defect detection of photovoltaic components in the prior art, large computing power usage and low detection probability, and the existence of a lot of redundant computing power and storage usage.
[0004] The present application provides a photovoltaic defect remote monitoring method based on wireless communication transmission, the method comprising: when a photovoltaic component to be inspected enters a first area of a carrier, activating a first tracking monitoring component to follow and shoot the photovoltaic component to be inspected at the same speed according to a first acquisition mode, and generating a grayscale feature image of the photovoltaic component; transmitting the grayscale feature image of the photovoltaic component to a cloud service center through wireless communication for image segmentation, and obtaining a quasi-defect identification area; transmitting the quasi-defect identification area to a second tracking monitoring component through wireless communication, and when the photovoltaic component to be inspected enters a second area of a carrier, activating the second tracking monitoring component to follow and shoot the quasi-defect identification area of the photovoltaic component to be inspected at the same speed according to a second acquisition mode, and obtaining a rich feature image of the photovoltaic component; transmitting the rich feature image of the photovoltaic component to the cloud service center through wireless communication for defect identification, and obtaining a defect detection result; marking the photovoltaic component to be inspected according to the defect detection result, generating a defect detection mark, and transmitting the defect detection mark to a user end of the cloud service center through wireless communication.
[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0006] The photovoltaic defect remote monitoring method based on wireless communication transmission provided by the present application relates to the field of wireless communication technology. A first tracking and monitoring component is used to perform same-speed tracking and shooting to generate a photovoltaic component grayscale feature image, which is transmitted to a cloud service center for image segmentation to obtain a quasi-defect identification area, and a second tracking and monitoring component is used to perform same-speed tracking and shooting of the quasi-defect identification area to obtain a rich feature image of the photovoltaic component, which is transmitted to a cloud service center for defect identification to obtain a defect detection result. The method solves the technical problems of small-scale defect detection of photovoltaic components in the prior art, large computing power occupancy and low detection probability, and the existence of a large amount of redundant computing power and storage occupancy, and realizes the technical effect of remote monitoring of photovoltaic defects through wireless communication transmission technology, improving the probability of defect detection of photovoltaic components, and reducing computing power redundancy. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0008] Figure 1 A schematic diagram of a process flow of a photovoltaic defect remote monitoring method based on wireless communication transmission provided in an embodiment of the present application;
[0009] Figure 2 A schematic diagram of a process for obtaining a quasi-defect identification area in a photovoltaic defect remote monitoring method based on wireless communication transmission provided in an embodiment of the present application;
[0010] Figure 3 A schematic diagram of a process for obtaining defect detection results in a photovoltaic defect remote monitoring method based on wireless communication transmission provided in an embodiment of the present application. DETAILED DESCRIPTION
[0011] The present application provides a photovoltaic defect remote monitoring method based on wireless communication transmission, which is used to solve the technical problems of small-scale defect detection of photovoltaic components in the prior art, large computing power usage and low detection probability, and the existence of a lot of redundant computing power and storage usage.
[0012] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0013] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices.
[0014] Embodiment 1:
[0015] like Figure 1 As shown, the present application provides a photovoltaic defect remote monitoring method based on wireless communication transmission, which is applied to a photovoltaic defect remote monitoring system based on wireless communication transmission. The system includes a cloud service center, and the cloud service center is communicatively connected with a first tracking monitoring component and a second tracking monitoring component. The method includes:
[0016] P10: When the photovoltaic component to be inspected enters the first area of the carrier, the first tracking monitoring component is activated to perform the same-speed tracking and photographing of the photovoltaic component to be inspected in a first acquisition mode to generate a grayscale feature image of the photovoltaic component.
[0017] Specifically, a photoelectric sensor, visual recognition or other sensor is used to detect whether the photovoltaic component to be inspected has entered the first area of the carrier, and the first area of the carrier is usually pre-set according to the transmission path and detection requirements of the photovoltaic component. Once it is detected that the photovoltaic component to be inspected has entered the first area, the first tracking monitoring component is activated. The first tracking monitoring component is specially designed to follow and photograph the photovoltaic component, and may include one or more cameras, image sensors, motion control systems, etc.
[0018] Furthermore, the first tracking and monitoring component is used to track and shoot the photovoltaic component to be inspected at the same speed according to the first acquisition mode, that is, to ensure that the movement speed of the camera or image sensor of the first tracking and monitoring component is completely consistent with the transmission speed of the photovoltaic component on the carrier, and to track and shoot the photovoltaic component to be inspected at the same speed to ensure that the captured image is clear and stable, and will not be blurred or distorted due to the movement of the photovoltaic component. Among them, the first acquisition mode includes the settings of parameters such as image resolution, frame rate, exposure time, white balance, etc., and the selection of these parameters depends on the specific detection requirements and scene conditions.
[0019] After the shooting is completed, the captured original image data (which may be a color image) is converted into a photovoltaic module grayscale feature image. The photovoltaic module grayscale feature image only contains brightness information, not color information, so it can simplify the subsequent image processing and analysis process. At the same time, the photovoltaic module grayscale feature image can also highlight the texture, shadow and other features of the photovoltaic module surface, which is helpful for subsequent defect detection.
[0020] Furthermore, before step P10 in the embodiment of the present application, step P10a is also included. Step P10a includes:
[0021] P11a: obtaining the movement speed timing information of the photovoltaic assembly to be inspected in the first area of the carrier according to the carrier control parameters;
[0022] P12a: Initialize the motion parameters of the first tracking monitoring component according to the motion speed timing information.
[0023] It should be understood that in order to ensure that the first tracking and monitoring component can accurately track and shoot according to the actual movement of the photovoltaic component to be inspected, the carrier control parameters, including the driving speed, acceleration, deceleration, etc. of the carrier, are referred to to calculate the movement speed timing information of the photovoltaic component to be inspected in the first area of the carrier, and the movement speed timing information describes the change of the speed of the photovoltaic component over time when it passes through the first area. Further, according to the movement speed timing information, the movement parameters of the first tracking and monitoring component are initialized and set, and the movement parameters include the moving speed, acceleration, deceleration, etc. of the camera, as well as the image acquisition parameters such as focus and exposure of the camera. Through the initialization of the movement parameters, the first tracking and monitoring component can accurately adjust its own movement state to match the movement speed of the photovoltaic component to be inspected, thereby ensuring that the captured image is clear and stable, and can accurately reflect the actual state of the photovoltaic component.
[0024] P20: Transmitting the photovoltaic module grayscale feature image to a cloud service center via wireless communication for image segmentation to obtain a quasi-defect identification area.
[0025] Further, such as Figure 2 As shown, step P20 of the embodiment of the present application also includes:
[0026] P21: Obtaining a reference grayscale image of a photovoltaic component, wherein the reference grayscale image of the photovoltaic component is generated by photographing a healthy photovoltaic component by the first tracking monitoring component according to the first acquisition mode;
[0027] P22: performing pixel-level comparison on the photovoltaic module grayscale feature image and the photovoltaic module reference grayscale image to generate a grayscale deviation matrix, wherein the grayscale deviation positions of the grayscale deviation matrix correspond to the pixel coordinates of the photovoltaic module grayscale feature image in a one-to-one manner;
[0028] P23: extracting a quasi-defective pixel coordinate set whose grayscale deviation is greater than or equal to a grayscale deviation threshold according to the grayscale deviation matrix;
[0029] P24: Segment the photovoltaic module grayscale feature image according to the quasi-defect pixel coordinate set to generate the quasi-defect identification area.
[0030] Optionally, the photovoltaic module grayscale feature image is transmitted to a cloud service center via wireless communication, and the cloud service center performs image segmentation to identify potential defect areas, i.e., quasi-defect identification areas. Specifically, a reference grayscale image of the photovoltaic module is obtained, i.e., an image generated by capturing a healthy photovoltaic module by the first tracking monitoring component in accordance with the first acquisition mode, which represents the grayscale characteristics of the photovoltaic module in a defect-free state. Furthermore, the received photovoltaic module grayscale feature image is compared pixel by pixel with the reference grayscale image. A grayscale deviation matrix is generated by calculating the grayscale value difference of each pixel. Each grayscale deviation position in the grayscale deviation matrix corresponds one-to-one to a pixel coordinate in the grayscale feature image of the photovoltaic module.
[0031] Furthermore, based on the defect sample characteristics, a grayscale deviation threshold is set to determine the pixel points with grayscale deviation caused by the defect. By comparing each element in the grayscale deviation matrix with the grayscale deviation threshold, a pixel coordinate set with grayscale deviation greater than or equal to the grayscale deviation threshold is extracted as a quasi-defective pixel coordinate set, representing a possible defective area.
[0032] Furthermore, the photovoltaic module grayscale feature image is segmented according to the quasi-defective pixel coordinate set, for example, by marking the pixel area represented by the quasi-defective pixel coordinate set to form a quasi-defect identification area, providing reference information for subsequent detailed detection and defect identification.
[0033] Furthermore, before step P23 in the embodiment of the present application, step P23a is also included. Step P23a includes:
[0034] P23-1a: photographing the defective sample photovoltaic components according to the first acquisition mode by the first tracking monitoring component to obtain a set of defective sample grayscale images;
[0035] P23-2a: Calculate the defect area grayscale of the defect sample grayscale image set and the defect sample grayscale deviation set of the photovoltaic module reference grayscale image;
[0036] P23-3a: Extract the minimum value of the grayscale deviation set of the defective samples and set it as the grayscale deviation threshold.
[0037] Specifically, before extracting the quasi-defective pixel coordinate set, it is necessary to determine the grayscale deviation threshold. First, use the first tracking monitoring component to shoot the known defective sample photovoltaic components according to the set first acquisition mode. The defective sample photovoltaic components should have defects of different types and degrees, so as to obtain rich defect sample data. Through shooting, a defective sample grayscale image set is obtained.
[0038] Furthermore, the defective area in each defective sample grayscale image is compared with the reference grayscale image of the photovoltaic module, and a defective sample grayscale deviation set is obtained by calculating the grayscale deviation between the defective area and the corresponding area of the reference image. The defective sample grayscale deviation set contains the grayscale deviation values of the defective areas in all defective sample images. Furthermore, the minimum value of the defective sample grayscale deviation set is extracted and set as the grayscale deviation threshold. The grayscale deviation threshold can be used to determine whether a pixel belongs to a defective area. If the grayscale deviation value of a pixel is greater than or equal to the grayscale deviation threshold, it is considered that the pixel may belong to a defective area.
[0039] P30: The quasi-defect identification area is transmitted to the second tracking monitoring component via wireless communication. When the photovoltaic component to be inspected enters the second area of the carrier, the second tracking monitoring component is activated to perform same-speed tracking and shooting of the quasi-defect identification area of the photovoltaic component to be inspected in accordance with the second acquisition mode to obtain a rich feature image of the photovoltaic component.
[0040] It should be understood that after the cloud service center obtains the quasi-defect identification area through image segmentation, the information of the quasi-defect identification area will be transmitted to the second tracking monitoring component through wireless communication. The second tracking monitoring component is a device specially designed for more detailed detection of photovoltaic modules, with higher image resolution and more accurate motion control system.
[0041] When the photovoltaic module to be inspected moves from the first area to the second area of the carrier, the second tracking monitoring component is activated. After receiving the information of the quasi-defect identification area, the second tracking monitoring component follows and photographs the quasi-defect identification area at the same speed according to the pre-set second acquisition mode. The second acquisition mode may include higher image resolution, different spectral ranges, more accurate exposure and focus settings, etc.
[0042] By following and shooting at the same speed, the second following monitoring component can capture the detailed features of the photovoltaic component to be inspected in the quasi-defect identification area and generate a rich feature image of the photovoltaic component. The rich feature image of the photovoltaic component contains not only grayscale information, but also spectral information, texture information, shape information, etc., providing rich data support for subsequent defect identification and analysis.
[0043] Furthermore, before step P30 in the embodiment of the present application, step P30a is also included. Step P30a includes:
[0044] P31a: Determine whether the number of regions of the quasi-defect identification region is equal to 0;
[0045] P32a: If the number of the quasi-defect identification areas is equal to 0, the second tracking monitoring component is not activated, and the surface of the photovoltaic component to be inspected is marked as defect-free;
[0046] P33a: If the number of the quasi-defect identification areas is not equal to 0, when the photovoltaic component to be inspected enters the second area of the carrier, the second tracking monitoring component is activated to perform the same-speed tracking and shooting of the quasi-defect identification areas of the photovoltaic component to be inspected, so as to obtain a rich feature image of the photovoltaic component.
[0047] Optionally, before activating the second tracking monitoring component, it is necessary to determine whether a quasi-defect identification area exists. First, the number of generated quasi-defect identification areas will be determined. If the number of areas in the quasi-defect identification area is equal to 0, it means that no potential defects were found on the surface of the photovoltaic component to be inspected in the preliminary inspection. In this case, the second tracking monitoring component may not be activated for a secondary detailed inspection, but the photovoltaic component may be directly marked as defect-free on the surface to reduce the amount of calculation. If the number of areas in the quasi-defect identification area is not equal to 0, it means that a possible defective area was found in the preliminary inspection. In this case, when the photovoltaic component to be inspected enters the second area of the carrier, the second tracking monitoring component is activated, and the photovoltaic component is tracked and photographed at the same speed based on the information of the quasi-defect identification area to obtain richer and more detailed characteristic images of the photovoltaic component for subsequent defect identification and analysis.
[0048] P40: Transmitting the photovoltaic module rich feature image to the cloud service center via wireless communication for defect identification to obtain defect detection results.
[0049] Further, such as Figure 3 As shown, step P40 of the embodiment of the present application also includes:
[0050] P41: Obtaining a photovoltaic component benchmark rich feature image, wherein the photovoltaic component benchmark rich feature image is generated by photographing a healthy photovoltaic component by the second tracking monitoring component according to the second acquisition mode;
[0051] P42: extracting features from the photovoltaic module rich feature image to obtain a first feature extraction result;
[0052] P43: extracting features from the photovoltaic module benchmark rich feature image to obtain a second feature extraction result;
[0053] P44: The first feature extraction result and the second feature extraction result are fused and analyzed by a defect classifier to generate the defect detection result.
[0054] It should be understood that the rich feature image of the photovoltaic component captured by the second tracking monitoring component is transmitted to the cloud service center and defect identification is performed. Before defect identification is performed, a baseline rich feature image of a healthy photovoltaic component is first obtained, and the baseline rich feature image is generated by the second tracking monitoring component shooting the healthy photovoltaic component according to the second acquisition mode, and represents the detailed features of the photovoltaic component in a defect-free state.
[0055] Furthermore, image processing technology is used to extract key features such as texture, shape, color, etc. from the photovoltaic module rich feature image, which are encoded and stored as the first feature extraction result. Similarly, feature extraction is performed on the photovoltaic module benchmark rich feature image to obtain the feature code of the healthy photovoltaic module, which is stored as the second feature extraction result.
[0056] Furthermore, the first feature extraction result and the second feature extraction result are fused and analyzed by a defect classifier, wherein the defect classifier is a pre-trained machine learning model that can identify whether a photovoltaic module has defects based on the input feature extraction result. In this step, the defect classifier fuses and analyzes the first feature extraction result (i.e., the feature of the photovoltaic module to be inspected) and the second feature extraction result (i.e., the feature of a healthy photovoltaic module), and generates a defect detection result by comparing the difference between the two. The defect detection result can reflect whether the photovoltaic module to be inspected has defects, as well as the type and degree of the defects.
[0057] Furthermore, step P42 of the embodiment of the present application also includes:
[0058] P42-1: Extracting texture features from the photovoltaic module rich feature image to obtain a texture feature extraction result;
[0059] P42-2: performing shape feature extraction on the photovoltaic module rich feature image to obtain a shape feature extraction result;
[0060] P42-3: Extracting size features from the photovoltaic module rich feature image to obtain a size feature extraction result;
[0061] P42-4: Performing color feature extraction on the photovoltaic module rich feature image to obtain a color feature extraction result;
[0062] P42-5: Add the texture feature extraction result, the shape feature extraction result, the size feature extraction result and the color feature extraction result into the first feature extraction result.
[0063] Specifically, the image processing algorithm is used to extract texture features from the rich feature images of the photovoltaic module. The texture features describe the spatial distribution and arrangement rules between pixels in the image, and have a high description capability for the fine structure and material characteristics of the surface of the photovoltaic module. At the same time, the shape features of the rich feature images of the photovoltaic module are extracted, including extracting shape features such as edge detection, contour extraction, and corner point detection, as important parameters for describing the appearance contour and geometric shape of the object. Further, the size features of the rich feature images of the photovoltaic module are extracted, such as measuring the length, width, area, etc. of the defects, which can be used to identify defects with a specific size range, such as cracks, scratches, etc. At the same time, the color features of the rich feature images of the photovoltaic module are extracted, such as counting the number of pixels in different color areas, calculating the color distribution histogram, etc., which are used to identify defects with abnormal colors, such as stains, discoloration, etc. Finally, the texture feature extraction results, the shape feature extraction results, the size feature extraction results, and the color feature extraction results are integrated to generate the first feature extraction result, which is used as input data for subsequent defect recognition and is used for comparison and analysis with the benchmark feature extraction results.
[0064] Furthermore, step P44 of the embodiment of the present application also includes:
[0065] P44-1: Compare the first feature extraction result and the second feature extraction result to construct a feature deviation vector matrix;
[0066] P44-2: The defect classifier includes a plurality of sub-classification trees, wherein the plurality of sub-classification trees cannot be divided evenly by a preset number of division categories;
[0067] P44-3: Through the plurality of sub-classification trees, the characteristic deviation vector matrices are fitted respectively to generate a plurality of defect classification results;
[0068] P44-4: Fit the type with the maximum frequency of the several defect classification results and set it as the defect detection result.
[0069] Optionally, the process of fusion analysis of the feature extraction results may be to first compare the feature differences between the rich feature image of the photovoltaic component to be inspected (first feature extraction result) and the baseline feature image of the healthy photovoltaic component (second feature extraction result), including comparing the differences between multiple features such as texture, shape, size and color, and calculating the differences between each feature to combine into a feature deviation vector matrix, which can quantitatively describe the degree of difference between the photovoltaic component to be inspected and the healthy photovoltaic component.
[0070] Furthermore, a defect classifier is used for feature fusion. Unlike a traditional single classifier, the defect classifier includes several sub-classification trees. Multiple sub-classification trees can improve the accuracy and robustness of classification, and the number of pre-set classification categories of the several sub-classification trees cannot be divided evenly, which means that each sub-classification tree may be responsible for identifying multiple types of defects, or certain types of defects may be jointly identified by multiple sub-classification trees. This design increases the flexibility and adaptability of the classifier. Optionally, the feature deviation vector matrix is input into each sub-classification tree for fitting and classification. Each sub-classification tree analyzes and judges the feature deviation vector matrix according to its own training data and classification rules to generate a defect classification result. Since there are multiple sub-classification trees, several defect classification results will be generated.
[0071] Furthermore, the above-mentioned defect classification results are counted and fused, the frequency of each defect type in the classification results of all sub-classification trees is calculated, and the defect type with the highest frequency is set as the final defect detection result. This frequency-based fusion method can comprehensively consider the classification results of multiple sub-classification trees and improve the accuracy and reliability of defect detection. At the same time, it improves the flexibility and adaptability of the system, so that it can adapt to different types and degrees of defect detection needs.
[0072] P50: Identify the photovoltaic assembly to be inspected according to the defect detection result, generate a defect detection identifier, and transmit the defect detection identifier to the user end of the cloud service center through wireless communication.
[0073] It should be understood that the photovoltaic module to be inspected is identified according to the defect detection result. If the photovoltaic module to be inspected is detected to have defects, a defect detection mark is generated, which includes information such as the type, location, and degree of the defect. If the photovoltaic module to be inspected is detected to be defect-free, a defect-free mark is generated. And it is transmitted to the user end of the cloud service center by wireless communication. This user end may be a platform or application specifically used to receive and process the inspection results of photovoltaic modules. Through wireless communication transmission, users can understand the inspection results of the photovoltaic modules to be inspected in real time, and take corresponding measures as needed to ensure the normal operation and power generation efficiency of the photovoltaic system.
[0074] In summary, the embodiments of the present application have at least the following technical effects:
[0075] The present application uses a first tracking and monitoring component to perform same-speed tracking and shooting to generate a grayscale feature image of the photovoltaic component, which is transmitted to a cloud service center for image segmentation to obtain a quasi-defect identification area, and uses a second tracking and monitoring component to perform same-speed tracking and shooting of the quasi-defect identification area to obtain a rich feature image of the photovoltaic component, which is transmitted to a cloud service center for defect identification to obtain a defect detection result.
[0076] The technical effect of remotely monitoring photovoltaic defects through wireless communication transmission technology, improving the probability of detecting defects in photovoltaic modules, and reducing computing power redundancy has been achieved.
[0077] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0078] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.
[0079] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.
Claims
1. A photovoltaic defect remote monitoring method based on wireless communication transmission, characterized in that: A photovoltaic defect remote monitoring system applied to wireless communication transmission, the system includes a cloud service center, the cloud service center is communicatively connected with a first tracking monitoring component and a second tracking monitoring component, and includes: When the photovoltaic component to be inspected enters the first area of the carrier, the first tracking monitoring component is activated to follow and photograph the photovoltaic component to be inspected at the same speed according to the first acquisition mode to generate a grayscale feature image of the photovoltaic component; Transmitting the photovoltaic module grayscale feature image to a cloud service center via wireless communication for image segmentation to obtain a quasi-defect identification area; The quasi-defect identification area is transmitted to the second tracking monitoring component through wireless communication, and when the photovoltaic component to be inspected enters the second area of the carrier, the second tracking monitoring component is activated to perform the same-speed tracking and shooting of the quasi-defect identification area of the photovoltaic component to be inspected in accordance with the second acquisition mode, so as to obtain a rich feature image of the photovoltaic component; Transmitting the photovoltaic module rich feature image to the cloud service center via wireless communication for defect identification to obtain defect detection results; Identify the photovoltaic assembly to be inspected according to the defect detection result, generate a defect detection identifier, and transmit the defect detection identifier to a user end of the cloud service center via wireless communication; The photovoltaic module grayscale feature image is transmitted to the cloud service center through wireless communication for image segmentation to obtain a quasi-defect identification area, including: Obtaining a photovoltaic component reference grayscale image, wherein the photovoltaic component reference grayscale image is generated by photographing a healthy photovoltaic component by the first tracking monitoring component according to the first acquisition mode; Performing pixel-level comparison on the photovoltaic module grayscale feature image and the photovoltaic module reference grayscale image to generate a grayscale deviation matrix, wherein the grayscale deviation positions of the grayscale deviation matrix correspond one-to-one to the pixel coordinates of the photovoltaic module grayscale feature image; According to the grayscale deviation matrix, extracting a quasi-defective pixel coordinate set whose grayscale deviation is greater than or equal to a grayscale deviation threshold; The photovoltaic module grayscale feature image is segmented according to the quasi-defect pixel coordinate set to generate the quasi-defect identification area.
2. The photovoltaic defect remote monitoring method based on wireless communication transmission according to claim 1, characterized in that: The quasi-defect identification area is transmitted to a second tracking monitoring component through wireless communication. When the photovoltaic component to be inspected enters the second area of the carrier, the second tracking monitoring component is activated to perform same-speed tracking and shooting of the quasi-defect identification area of the photovoltaic component to be inspected to obtain a rich feature image of the photovoltaic component. The method also includes: Determining whether the number of regions of the quasi-defect identification region is equal to 0; If the number of the quasi-defect identification areas is equal to 0, the second tracking monitoring component is not activated, and the photovoltaic component to be inspected is marked as having no surface defects; If the number of the quasi-defect identification areas is not equal to 0, when the photovoltaic component to be inspected enters the second area of the carrier, the second tracking monitoring component is activated to track and shoot the quasi-defect identification areas of the photovoltaic component to be inspected at the same speed to obtain a rich feature image of the photovoltaic component.
3. The photovoltaic defect remote monitoring method based on wireless communication transmission according to claim 1, characterized in that: According to the grayscale deviation matrix, a quasi-defective pixel coordinate set whose grayscale deviation is greater than or equal to a grayscale deviation threshold is extracted, wherein the method further comprises: Photographing the defective sample photovoltaic components according to the first acquisition mode by the first tracking monitoring component to obtain a set of defective sample grayscale images; Calculating a defect area grayscale of the defect sample grayscale image set and a defect sample grayscale deviation set of the photovoltaic module reference grayscale image; The minimum value of the grayscale deviation set of the defective samples is extracted and set as the grayscale deviation threshold.
4. The photovoltaic defect remote monitoring method based on wireless communication transmission according to claim 1, characterized in that: The photovoltaic module rich feature image is transmitted to the cloud service center through wireless communication for defect identification to obtain defect detection results, including: Obtaining a photovoltaic component benchmark rich feature image, wherein the photovoltaic component benchmark rich feature image is generated by photographing a healthy photovoltaic component by the second tracking monitoring component according to the second acquisition mode; Performing feature extraction on the photovoltaic component rich feature image to obtain a first feature extraction result; Performing feature extraction on the photovoltaic module benchmark rich feature image to obtain a second feature extraction result; The first feature extraction result and the second feature extraction result are fused and analyzed by a defect classifier to generate the defect detection result.
5. The photovoltaic defect remote monitoring method based on wireless communication transmission according to claim 4, characterized in that: Performing feature extraction on the photovoltaic component rich feature image to obtain a first feature extraction result includes: Extracting texture features from the photovoltaic module rich feature image to obtain a texture feature extraction result; Performing shape feature extraction on the photovoltaic component rich feature image to obtain a shape feature extraction result; Extracting size features from the photovoltaic module rich feature image to obtain a size feature extraction result; Performing color feature extraction on the photovoltaic component rich feature image to obtain a color feature extraction result; The texture feature extraction result, the shape feature extraction result, the size feature extraction result and the color feature extraction result are added to the first feature extraction result.
6. The photovoltaic defect remote monitoring method based on wireless communication transmission according to claim 4, characterized in that: The defect classifier performs a fusion analysis on the first feature extraction result and the second feature extraction result to generate the defect detection result, including: Comparing the first feature extraction result and the second feature extraction result to construct a feature deviation vector matrix; The defect classifier includes a plurality of sub-classification trees, wherein the plurality of sub-classification trees cannot be divided evenly by a preset number of division categories; Through the plurality of sub-classification trees, the characteristic deviation vector matrices are respectively fitted to generate a plurality of defect classification results; The type with the largest frequency of the plurality of defect classification results is fitted and set as the defect detection result.
7. The photovoltaic defect remote monitoring method based on wireless communication transmission according to claim 1, characterized in that: When the photovoltaic component to be inspected enters the first area of the carrier, the first tracking monitoring component is activated to follow and shoot the photovoltaic component to be inspected at the same speed according to the first acquisition mode to generate a grayscale feature image of the photovoltaic component, which includes: According to the carrier control parameters, obtaining the movement speed timing information of the photovoltaic assembly to be inspected in the first area of the carrier; The motion parameters of the first tracking monitoring component are initialized according to the motion speed timing information.