Photovoltaic glass inspection method and system
The automated inspection of photovoltaic glass using a high-precision transport platform and convolutional neural networks addresses the limitations of manual and simple image processing methods, achieving precise and efficient defect detection and ensuring high-quality glass production.
Patent Information
- Application Number
- CN202411620510.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-13
- Publication Date
- 2025-03-14
AI Technical Summary
Current methods for inspecting photovoltaic glass rely on manual or simple image processing, leading to high subjectivity and low accuracy in detecting defects, particularly small flaws, resulting in substandard glass quality, reduced efficiency, and increased costs.
An automated inspection method involving a high-precision transport platform with camera-based imaging and convolutional neural networks to detect defects on the glass surface, edges, and floating dust, ensuring thorough and accurate classification of glass quality.
The method provides a robust, automated, and efficient inspection process that accurately identifies defects, ensuring high-quality glass production with reduced human intervention and increased precision, thereby enhancing production efficiency and product reliability.
Smart Images
Figure CN119619140A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of photovoltaic glass inspection, and in particular to a photovoltaic glass inspection method and system. Background Art
[0002] Photovoltaic glass is a special glass specifically used in solar photovoltaic power generation systems. Its main function is to protect solar cell modules while maximizing the transmittance of light, thereby improving the power generation efficiency of photovoltaic cells. The photovoltaic glass inspection method uses a glass inspection machine to use machine vision technology, optical sensing technology and image processing algorithms to identify and classify defects on the glass surface, such as bubbles, stones, scratches and color spots.
[0003] In photovoltaic power generation systems, photovoltaic glass covers the surface of solar panels to protect solar cells. Therefore, photovoltaic glass needs to undergo strict quality inspections during the production process to ensure that its light transmittance, surface cleanliness, thickness uniformity, and edge defects meet the standards. Therefore, strict inspection of photovoltaic glass is a key step in ensuring the quality of photovoltaic cell modules, which is of great significance to reducing production and operation and maintenance costs, enhancing the competitiveness of corporate products, and quality control of the photovoltaic industry chain.
[0004] Traditional photovoltaic glass inspection methods rely on manual visual inspection or simple sensors and image processing methods, which are highly subjective and have low defect detection accuracy. For some minor defects (such as fine cracks, tiny particles, surface contamination, etc.), manual or simple equipment may not be able to accurately detect them, resulting in substandard photovoltaic glass quality, low efficiency, low automation level and low detection accuracy, which affects production efficiency and increases labor costs. Summary of the invention
[0005] A technical problem to be solved by the present application is that the existing photovoltaic glass inspection method relies on manual visual inspection or simple sensors and image processing methods. This method is highly subjective and has low defect detection accuracy. For some minor defects, manual or simple equipment may not be able to accurately detect them, resulting in substandard quality of photovoltaic glass, low efficiency, low level of automation and low detection accuracy, which affects production efficiency and increases labor costs.
[0006] In order to solve the above technical problems, the embodiment of the present application provides a photovoltaic glass inspection method, comprising:
[0007] S1: Place the glass to be inspected on a high-precision transport platform;
[0008] S2: When the high-precision transport platform is transporting the glass to be inspected, the camera is used to capture the image of the glass surface;
[0009] S3: According to the glass surface image, the defects of the glass to be inspected are detected to determine whether the glass to be inspected has defects; if so, the glass to be inspected is discarded; otherwise, proceed to the next step;
[0010] S4: Take images of floating dust on the glass surface by using a camera;
[0011] S5: According to the dust image on the glass surface, the defects of the glass to be inspected are detected to determine whether the glass to be inspected has defects; if so, the glass to be inspected is discarded; otherwise, proceed to the next step;
[0012] S6: When the high-precision transport platform is transporting the glass to be inspected, the camera is used to capture the image of the glass edge;
[0013] S7: According to the glass edge image, the defects of the glass to be inspected are inspected to determine whether the glass to be inspected has defects; if so, the glass to be inspected is discarded; otherwise, the glass to be inspected is determined to be qualified.
[0014] In some embodiments, S1 specifically includes:
[0015] The glass to be inspected is turned over by a turning machine, and is placed on a high-precision transportation platform.
[0016] In some embodiments, the high-precision transport platform specifically includes: transport rollers and air flotation strips;
[0017] The transport roller is used to make the glass move laterally;
[0018] The air flotation strip is used to float the inclined glass by high-pressure air.
[0019] In some embodiments, the defect types detected by the glass surface image include: bubbles, scratches, color spots and cracks.
[0020] In some embodiments, the step S3 of detecting defects of the glass to be detected according to the glass surface image to determine whether the glass to be detected has defects specifically includes:
[0021] S301: Decomposing the glass surface image into an R channel image, a G channel image and a B channel image;
[0022] S302: Inputting the R channel image, the G channel image and the B channel image into a convolutional neural network;
[0023] S303: performing feature extraction on the input R channel image, the G channel image and the B channel image respectively through the convolutional layers in the convolutional neural network:
[0024] F1=ReLU(K1*I1+b1)
[0025] F2=ReLU(K2*I2+b2)
[0026] F3=ReLU(K3*I3+b3)
[0027] Wherein, F1, F2 and F3 represent the first feature map, the second feature map and the third feature map respectively, K1, K2 and K3 represent the convolution kernel of the R channel image, the convolution kernel of the G channel image and the convolution kernel of the B channel image respectively, ReLU represents the activation function, * represents the convolution operation, I1, I2 and I3 represent the input R channel image, G channel image and B channel image respectively, b1, b2 and b3 represent the bias term of the convolution kernel of the R channel image, the bias term of the convolution kernel of the G channel image and the bias term of the convolution kernel of the B channel image respectively;
[0028] S304: performing feature fusion on the first feature map, the second feature map, and the third feature map by weighted averaging:
[0029] F fused =w1·F1+w2·F2+w3·F3
[0030] Among them, F fused represents the fused feature map, w1, w2 and w3 represent the weight coefficient of the first feature map, the weight coefficient of the second feature map and the weight coefficient of the third feature map respectively;
[0031] S305: According to the fused feature map, based on the Softmax activation function, detect whether there is a defect, and take the defect type with the highest probability as the detection result:
[0032] P = Softmax(WF fused +b)
[0033] P=(p1,p2,…,p n )
[0034] Among them, P represents the defect detection result, p i represents the probability of the existence of the i-th type of defect, Softmax represents the Softmax activation function, W represents the weight matrix, F fused represents the fused feature map, and b represents the bias term.
[0035] In some embodiments, the glass surface dust image is used to detect defect types including: dust, dirt, oil stains, and water stains.
[0036] In some embodiments, the defect types detected by the glass edge image include: edge chipping, corner chipping, and edge irregularity.
[0037] In some embodiments, after S7, the method further includes:
[0038] The glass surface image, the glass surface dust image and the glass edge image are re-photographed by the camera, and the defects of the glass to be inspected are re-inspected based on the re-photographed glass surface image, the glass surface dust image and the glass edge image.
[0039] The embodiment of the present application provides a photovoltaic glass inspection system, comprising:
[0040] processor;
[0041] A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the photovoltaic glass inspection method as described above is implemented.
[0042] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the program implements the above-mentioned photovoltaic glass inspection method when executed by a processor.
[0043] Through the above technical solution, the glass to be inspected is placed on a high-precision transportation platform, and the front image of the glass is photographed during transportation to obtain surface information. The inspection is performed based on the photographed front image of the glass to determine whether there are significant defects. If defects are detected, the glass is discarded. During transportation, the floating dust image of the glass is continuously photographed, and the granular defects on the surface are detected by analyzing the floating dust image. If defects exist, the glass is discarded. Otherwise, the edge image of the glass is further photographed during transportation, and the defects, especially edge defects, are determined based on the edge image. If they exist, the glass is discarded; otherwise, the glass is confirmed to be qualified. This automated photovoltaic glass inspection method comprehensively inspects the surface, floating dust particles and edge defects of photovoltaic glass through a gradually refined inspection process to ensure product quality. At the same time, the system has a simple structure, a high degree of automation, high inspection efficiency, and stable operation. It can accurately determine the defect type and type inside the glass, and quickly determine whether the glass is good or bad, realizing fully automatic surface inspection, particle inspection, edge inspection, re-inspection and other actions on the glass. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0045] Figure 1 It is a schematic flow chart of a photovoltaic glass inspection method disclosed in an embodiment of the present application;
[0046] Figure 2 It is a process flow chart of a photovoltaic glass inspection method disclosed in an embodiment of the present application;
[0047] Figure 3 It is a structural schematic diagram of a photovoltaic glass inspection system disclosed in an embodiment of the present application. DETAILED DESCRIPTION
[0048] The following is a further detailed description of the implementation methods of the present application in conjunction with the accompanying drawings and examples. The detailed descriptions and drawings of the following examples are used to exemplarily illustrate the principles of the present application, but cannot be used to limit the scope of the present application. The present application can be implemented in many different forms and is not limited to the specific embodiments disclosed herein, but includes all technical solutions that fall within the scope of the claims.
[0049] The present application provides these embodiments to make the present application thorough and complete, and to fully express the scope of the present application to those skilled in the art. It should be noted that unless otherwise specifically stated, the relative arrangement of the parts and steps, the composition of the materials, the numerical expressions and the numerical values set forth in these embodiments should be interpreted as being merely exemplary, and not as limiting.
[0050] It should be noted that, in the description of this application, unless otherwise specified, the meaning of "multiple" is greater than or equal to two; the terms "upper", "lower", "left", "right", "inner", "outer", etc., indicating the orientation or positional relationship, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.
[0051] In addition, the words "first", "second" and similar words used in this application do not indicate any order, quantity or importance, but are only used to distinguish different parts. "Vertical" does not mean vertical in the strict sense, but is within the tolerance range. "Parallel" does not mean parallel in the strict sense, but is within the tolerance range. "Include" or "comprising" and similar words mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of including other elements.
[0052] It should also be noted that in the description of this application, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to the specific circumstances. When a specific device is described as being located between a first device and a second device, there may or may not be an intermediate device between the specific device and the first device or the second device.
[0053] All terms used in this application have the same meaning as those understood by those of ordinary skill in the art to which this application belongs, unless otherwise specifically defined. It should also be understood that terms defined in general dictionaries, such as general dictionaries, should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined herein.
[0054] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.
[0055] Figure 1 It is a flow chart of a photovoltaic glass inspection method disclosed in an embodiment of the present application.
[0056] Figure 2 It is a process flow chart of a photovoltaic glass inspection method disclosed in an embodiment of the present application.
[0057] The embodiment of the present application provides a glass product operation device and a photovoltaic glass inspection method, including:
[0058] S1: Place the glass to be inspected on the high-precision transport platform.
[0059] Among them, the glass to be inspected refers to photovoltaic glass that needs to be quality inspected, and the high-precision transportation platform refers to a mechanical device used to accurately transport glass, usually equipped with rollers, air flotation devices or precision transmission systems to ensure that the glass moves smoothly and accurately during the inspection process.
[0060] It should be noted that by placing the glass to be inspected on a high-precision transport platform, it can be ensured that the glass remains stable and flat during the entire inspection process, avoiding inspection errors caused by unstable glass movement. In addition, the use of a high-precision transport platform can realize automatic transportation of glass, improve inspection efficiency, reduce manual intervention, and reduce human errors in the production process.
[0061] In some embodiments, S1 specifically includes: turning over the glass to be inspected by a turning machine, and placing the glass to be inspected on a high-precision transportation platform.
[0062] In some embodiments, the high-precision transport platform specifically includes: transport rollers and air flotation strips. The transport rollers are used to make the glass move horizontally. The air flotation strips are used to float the tilted glass through high-pressure air. The tilted glass can float through high-pressure air to reduce the stress that the glass has to bear. The non-contact air flotation platform handling technology has the advantages of zero friction, zero loss, and high production speed.
[0063] Specifically, on the high-precision transportation platform, the glass to be inspected enters the first inspection process - the glass surface inspection process. In this process, the entire surface of the glass to be inspected is photographed by a high-resolution camera to capture the surface image to detect possible surface defects (such as scratches, bubbles, etc.). Then, the glass to be inspected continues to move to the next inspection point - the glass particle inspection process. In this process, a low-angle light source is used in conjunction with a camera to photograph the surface of the glass to be inspected, which is used to detect tiny pollutants such as floating dust particles and determine the cleanliness of the glass. Then, the glass to be inspected is transported to the glass edge inspection process, where the edge inspection camera is mainly used to photograph the four edges of the glass. The purpose of this process is to check the integrity of the glass edge, detect whether there are problems such as broken edges, missing corners or irregular edges, and complete the edge detection. After that, the glass to be inspected enters the glass re-inspection 1 process, which is mainly to confirm and classify the defects detected in the previous processes. This workstation is equipped with a camera module, which can efficiently confirm whether the defects exist and further refine the classification. If the glass passes the re-inspection 1 process, the glass to be inspected will continue to enter the glass re-inspection 2 process. At this workstation, high-precision detection equipment is used again to further confirm the defects to improve the accuracy and reliability of the detection. Finally, the glass to be inspected enters the glass re-inspection 3 process, which further detects and confirms the remaining minor defects. The re-inspection 3 process can further reduce the situation of false detection or missed detection. After all the re-inspection processes, the detection process ends. If the glass to be inspected is not judged as defective during the entire detection process, it is considered to be a qualified product.
[0064] S2: When the high-precision transport platform is transporting the glass to be inspected, the camera is used to capture the image of the glass surface.
[0065] Among them, the camera refers to an imaging device used to capture images of the photovoltaic glass surface, usually including a line scan camera or an area array camera, which is used to obtain glass surface information with high resolution. The glass surface image refers to the complete image of the photovoltaic glass surface captured by the camera, which is used to detect defects and irregularities on the glass surface.
[0066] Alternatively, the camera can use an industrial camera (CCD or CMOS) as an image acquisition device to capture the image of the target object through a lens. These cameras have the characteristics of high resolution and high speed, and can capture fast-moving targets.
[0067] Furthermore, in order to improve the contrast and clarity of the image, the visual inspection machine is equipped with a special lighting system. The lighting can be visible light or invisible light, including incandescent lamps, fluorescent lamps, LED lamps, etc. The lighting method may include backlight, front light, structured light or strobe light, etc.
[0068] Further, the collected analog image signals are converted into digital signals through an analog-to-digital converter (A / D) so that the computer can process them. The digital image contains information such as pixel distribution, brightness and color. The digital image is transmitted to the image processing system, and the system uses image processing algorithms to analyze the image. This may include edge detection, feature extraction, pattern recognition and other technologies to identify target features and defects in the image.
[0069] It should be noted that by taking images of the glass surface with a camera, accurate scanning and inspection of the glass surface can be achieved. Such a design can fully capture the details of the glass surface and effectively identify surface defects such as scratches, bubbles or other blemishes.
[0070] S3: According to the glass surface image, the defects of the glass to be inspected are detected to determine whether the glass to be inspected has defects. If so, the glass to be inspected is discarded. Otherwise, proceed to the next step.
[0071] It should be noted that by performing defect detection on the glass surface image, problematic glass can be identified and removed in a timely manner to ensure quality control of the production line. This step-by-step screening method can prevent defective glass from continuing to enter subsequent processes, thereby improving overall detection efficiency and accuracy.
[0072] Furthermore, the glass surface inspection process is mainly to conduct a preliminary inspection of the internal defects of the glass. This station is equipped with several line scan cameras, and three light sources are set on both sides of the camera and the back of the glass to supplement the light for the shooting area. The entire large surface of the glass is scanned and photographed. The approximate size and specific coordinates of the defects are scanned.
[0073] In some embodiments, the defect types detected by the glass surface image include: bubbles, scratches, color spots and cracks.
[0074] In some embodiments, the step S3 detects defects of the glass to be inspected based on the glass surface image and determines whether the glass to be inspected has defects, specifically including:
[0075] S301: Decomposing the glass surface image into an R channel image, a G channel image and a B channel image.
[0076] S302: Input the R channel image, the G channel image and the B channel image into the convolutional neural network.
[0077] S303: extracting features of the input R channel image, G channel image and B channel image respectively through the convolutional layers in the convolutional neural network:
[0078] F1=ReLU(K1*I1+b1)
[0079] F2=ReLU(K2*I2+b2)
[0080] F3=ReLU(K3*I3+b3)
[0081] Among them, F1, F2 and F3 represent the first feature map, the second feature map and the third feature map respectively, K1, K2 and K3 represent the convolution kernel of the R channel image, the convolution kernel of the G channel image and the convolution kernel of the B channel image respectively, ReLU represents the activation function, * represents the convolution operation, I1, I2 and I3 represent the input R channel image, G channel image and B channel image respectively, b1, b2 and b3 represent the bias term of the convolution kernel of the R channel image, the bias term of the convolution kernel of the G channel image and the bias term of the convolution kernel of the B channel image respectively.
[0082] S304: Perform feature fusion on the first feature map, the second feature map, and the third feature map by weighted averaging:
[0083] F fused =w1·F1+w2·F2+w3·F3
[0084] Among them, F fused represents the fused feature map, w1, w2 and w3 represent the weight coefficient of the first feature map, the weight coefficient of the second feature map and the weight coefficient of the third feature map respectively.
[0085] S305: According to the fused feature map, based on the Softmax activation function, detect whether there is a defect, and take the defect type with the highest probability as the detection result:
[0086] P = Softmax(WF fused +b)
[0087] P=(p1,p2,…,p n )
[0088] Among them, P represents the defect detection result, p i represents the probability of the existence of the i-th type of defect, Softmax represents the Softmax activation function, W represents the weight matrix, Ffused represents the fused feature map, and b represents the bias term.
[0089] S4: Use the camera to capture images of dust on the glass surface.
[0090] Among them, the image of floating dust on the glass surface is taken by a camera, which is specially used to detect images of floating dust particles or fine pollutants on the glass surface. It is shot through a specific light source and angle to clearly reflect the tiny particles and pollution.
[0091] S5: According to the dust image on the glass surface, the defects of the glass to be inspected are detected to determine whether the glass to be inspected has defects. If so, the glass to be inspected is discarded. Otherwise, proceed to the next step.
[0092] It should be noted that by analyzing the image of floating dust on the glass surface for defect detection, tiny particles and pollutants can be accurately identified, effectively ensuring the transmittance and cleanliness of the glass. This refined inspection method supplements the previous large-scale surface inspection, which can capture more subtle defects and further improve the overall quality of photovoltaic glass. At the same time, by promptly removing unqualified glass, defective products can be prevented from entering the next inspection link or production step, reducing the processing burden of subsequent processes, thereby improving the overall efficiency of the production line and the product qualification rate.
[0093] Furthermore, the glass particle inspection process is mainly to conduct a preliminary inspection of the internal defects of the glass. This station is equipped with several line scan cameras, and three light sources are set on both sides of the camera and the back of the glass to supplement the light for the shooting area. The entire large surface of the glass is scanned and photographed. The approximate size and specific coordinates of the defects are scanned.
[0094] Specifically, the method for detecting defects of the glass to be detected based on the dust image on the glass surface can refer to the method for detecting defects of the glass to be detected based on the glass surface image mentioned above, and the convolutional neural network is also used for detection. To avoid repetition, the present invention will not go into details.
[0095] In some embodiments, the defect types detected by the dust image on the glass surface include: dust, dirt, oil stains, and water stains.
[0096] S6: When the high-precision transport platform is transporting the glass to be inspected, the camera is used to capture the image of the glass edge.
[0097] Among them, the glass edge image refers to the image of the edge of the photovoltaic glass taken by a camera, which is used to detect the physical integrity of the glass edge. Edge detection aims to identify defects such as chipped corners, broken edges or cracks that may affect the strength and safety of the glass.
[0098] S7: According to the glass edge image, the defects of the glass to be inspected are inspected to determine whether the glass to be inspected has defects. If so, the glass to be inspected is discarded. Otherwise, the glass to be inspected is determined to be qualified.
[0099] It should be noted that by detecting the edge image of the glass, physical defects at the edge, such as missing corners, broken edges or cracks, can be effectively discovered. These defects usually affect the structural integrity and safety performance of the glass, especially during transportation, installation and long-term use, which may cause breakage. Timely detection and removal of glass with edge defects can improve the overall safety of photovoltaic modules, reduce the risk of breakage, and ensure the stability and service life of the final product. Compared with simple surface inspection, adding an edge detection link can comprehensively improve the quality control of photovoltaic glass and enhance the market competitiveness of the product.
[0100] Furthermore, the glass edge inspection process is equipped with four edge inspection cameras, which are mainly used to photograph the four edges of the glass to detect the grinding condition of the glass edges and whether there are defects such as chipped corners and broken edges.
[0101] In some embodiments, the defect types detected by the glass edge image include edge chipping, corner chipping, and edge irregularity.
[0102] Specifically, the method for detecting defects in the glass to be detected based on the glass edge image can refer to the method for detecting defects in the glass to be detected based on the glass surface image mentioned above, and a convolutional neural network is also used for detection. To avoid repetition, the present invention will not go into details.
[0103] In some embodiments, after S7, the method further includes:
[0104] The glass surface image, the glass surface dust image and the glass edge image are re-photographed by the camera, and the defects of the glass to be inspected are re-inspected based on the re-photographed glass surface image, the glass surface dust image and the glass edge image.
[0105] It should be noted that the glass re-inspection process is to determine and classify the defects of the glass. A total of three stations are set up for the re-inspection process. The reason for setting up three processes is to have a faster beat to produce. The first and second processes are set up with a camera module respectively, and the third inspection process is set up with two camera modules to take better and faster pictures of defects. After the shooting is completed, it will be transmitted to the system, and the system will make a logical judgment to determine whether the product meets the preset quality standards.
[0106] Through the above technical solution, the glass to be inspected is placed on a high-precision transportation platform, and the front image of the glass is photographed during transportation to obtain surface information. The inspection is performed based on the photographed front image of the glass to determine whether there are significant defects. If defects are detected, the glass is discarded. During transportation, the floating dust image of the glass is continuously photographed, and the granular defects on the surface are detected by analyzing the floating dust image. If defects exist, the glass is discarded. Otherwise, the edge image of the glass is further photographed during transportation, and the defects, especially edge defects, are determined based on the edge image. If they exist, the glass is discarded. Otherwise, the glass is confirmed to be qualified. This automated photovoltaic glass inspection method comprehensively inspects the surface, floating dust particles and edge defects of photovoltaic glass through a gradually refined inspection process to ensure product quality. At the same time, this system has a simple structure, a high degree of automation, high inspection efficiency, and stable operation. It can accurately determine the defect type and type inside the glass, and quickly determine whether the glass is good or bad, and realize fully automatic surface inspection, particle inspection, edge inspection, re-inspection and other actions on the glass.
[0107] Figure 3 It is a structural schematic diagram of a photovoltaic glass inspection system disclosed in an embodiment of the present application.
[0108] The embodiment of the present application provides a photovoltaic glass inspection system 20, which is applied to the above photovoltaic glass inspection method, including:
[0109] Processor 201.
[0110] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201 , the photovoltaic glass inspection method of the method embodiment is implemented.
[0111] The photovoltaic glass inspection system 20 provided by the present invention can execute the above-mentioned photovoltaic glass inspection method and achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.
[0112] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the program implements the above-mentioned photovoltaic glass inspection method when executed by a processor.
[0113] A computer-readable storage medium provided by the present invention can implement the steps and effects of the photovoltaic glass inspection method of the above method embodiment. To avoid repetition, the present invention will not go into details.
[0114] So far, various embodiments of the present application have been described in detail. In order to avoid obscuring the concept of the present application, some details known in the art are not described. Based on the above description, those skilled in the art can fully understand how to implement the technical solution disclosed herein.
[0115] Although some specific embodiments of the present application have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present application. It should be understood by those skilled in the art that the above embodiments may be modified or some technical features may be replaced by equivalents without departing from the scope and spirit of the present application. In particular, the various technical features mentioned in the various embodiments may be combined in any manner as long as there is no structural conflict.
Claims
1. A photovoltaic glass inspection method, characterized in that: include: S1: Place the glass to be inspected on a high-precision transport platform; S2: In the process of transporting the glass to be inspected by the high-precision transport platform, taking an image of the glass surface by a camera; S3: According to the glass surface image, the defects of the glass to be detected are detected to determine whether the glass to be detected has defects; if so, the glass to be detected is discarded; otherwise, proceed to the next step; S4: Take images of floating dust on the glass surface by using a camera; S5: According to the dust image on the glass surface, the defects of the glass to be detected are detected to determine whether the glass to be detected has defects; if so, the glass to be detected is discarded; otherwise, the process proceeds to the next step; S6: In the process of transporting the glass to be inspected by the high-precision transport platform, taking an image of the edge of the glass by a camera; S7: According to the glass edge image, the defects of the glass to be inspected are inspected to determine whether the glass to be inspected has defects; if so, the glass to be inspected is discarded; otherwise, the glass to be inspected is determined to be qualified.
2. The photovoltaic glass inspection method according to claim 1, characterized in that: The S1 is specifically: The glass to be inspected is turned over by a turning machine, and is placed on a high-precision transportation platform.
3. The photovoltaic glass inspection method according to claim 1, characterized in that: The high-precision transport platform specifically includes: transport rollers and air flotation strips; The transport roller is used to make the glass move laterally; The air flotation strip is used to float the inclined glass by high-pressure air.
4. The photovoltaic glass inspection method according to claim 1, characterized in that: The defect types that the glass surface image is used to detect include: bubbles, scratches, color spots and cracks.
5. The photovoltaic glass inspection method according to claim 1, characterized in that: The step S3 of detecting defects of the glass to be detected according to the glass surface image to determine whether the glass to be detected has defects specifically includes: S301: Decomposing the glass surface image into an R channel image, a G channel image and a B channel image; S302: Inputting the R channel image, the G channel image and the B channel image into a convolutional neural network; S303: performing feature extraction on the input R channel image, the G channel image and the B channel image respectively through the convolutional layers in the convolutional neural network: F1=ReLU(K1*I1+b1) F2=ReLU(K2*I2+b2) F3=ReLU(K3*I3+b3) Wherein, F1, F2 and F3 represent the first feature map, the second feature map and the third feature map respectively, K1, K2 and K3 represent the convolution kernel of the R channel image, the convolution kernel of the G channel image and the convolution kernel of the B channel image respectively, ReLU represents the activation function, * represents the convolution operation, I1, I2 and I3 represent the input R channel image, G channel image and B channel image respectively, b1, b2 and b3 represent the bias term of the convolution kernel of the R channel image, the bias term of the convolution kernel of the G channel image and the bias term of the convolution kernel of the B channel image respectively; S304: performing feature fusion on the first feature map, the second feature map, and the third feature map by weighted averaging: <h2 style=";text-align:left;direction:ltr">F<h2 style=";text-align:left;direction:ltr"> fused <h2 style=";text-align:left;direction:ltr"> = w1 F1 + w2 F2 + w3 F3 Among them, F fused represents a fused feature map, w1, w2 and w3 represent the weight coefficient of the first feature map, the weight coefficient of the second feature map and the weight coefficient of the third feature map respectively; S305: According to the fused feature map, based on the Softmax activation function, detect whether there is a defect, and take the defect type with the highest probability as the detection result: P=Softmax(WF fused +b) P=(p1,p2,…,p n ) Among them, P represents the defect detection result, p i represents the probability of the existence of the i-th type of defect, Softmax represents the Softmax activation function, W represents the weight matrix, F fused represents the fused feature map, and b represents the bias term.
6. The photovoltaic glass inspection method according to claim 1, characterized in that: The defect types that the glass surface dust image is used to detect include: dust, dirt, oil stains and water stains.
7. The photovoltaic glass inspection method according to claim 1, characterized in that: The glass edge image is used to detect defect types including edge chipping, corner chipping, and edge irregularity.
8. The photovoltaic glass inspection method according to claim 1, characterized in that: After S7, the method further includes: The glass surface image, the glass surface dust image and the glass edge image are re-photographed by the camera, and the defects of the glass to be inspected are re-inspected based on the re-photographed glass surface image, the glass surface dust image and the glass edge image.
9. A photovoltaic glass inspection system, characterized in that: include: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the photovoltaic glass inspection method according to any one of claims 1 to 8 is implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the photovoltaic glass inspection method according to any one of claims 1 to 8 is implemented.
Citation Information
Cited By
Photovoltaic module solid waste recycling and crushing device
CN120190197A
Photovoltaic module solid waste recycling and crushing device
CN120190197B