Production line control method and system based on photovoltaic glass thickness detection

Abstract

The invention discloses a production line control method and system based on photovoltaic glass thickness detection, and relates to the technical field of data detection. The method comprises the following steps: acquiring spectral information of a to-be-detected glass surface in a target production line; determining a plurality of detection points on the surface of the to-be-detected glass, and determining light features and glass contour features corresponding to the detection points based on the spectral information; inputting the light features and the glass contour features into a glass thickness detection model to obtain the initial thickness of the to-be-detected glass at each detection point, and calculating the target thickness of the to-be-detected glass based on the initial thickness of each detection point; collecting production parameters of the to-be-detected glass, and calculating the production lead-out quantity of the to-be-detected glass in combination with the production parameters and the target thickness; and adjusting the glass extraction amount of the target production line according to the difference between the production extraction amount and the standard production extraction amount. By implementing the technical scheme provided by the invention, the effect of improving the accuracy of production line control based on photovoltaic glass thickness detection is achieved.

Description

Technical Field

[0001] The present application relates to the field of data detection technology, and specifically to a production line control method and system based on photovoltaic glass thickness detection. Background Art

[0002] With the continuous growth of global energy demand and the rapid development of renewable energy technology, the photovoltaic industry, especially the manufacturing and application of photovoltaic glass, has received great attention. As a key component of solar panels, the quality of photovoltaic glass directly affects the conversion efficiency and life cycle of the entire photovoltaic system. Therefore, ensuring the quality of photovoltaic glass, especially the accuracy of its thickness, is crucial to the production of high-efficiency and high-reliability photovoltaic products.

[0003] At present, the existing production line control method based on photovoltaic glass thickness detection uses manual measurement and tools such as thickness gauges to measure the thickness at fixed edge positions on the glass. However, in actual applications, since the glass to be detected may have defects at different positions, it is easy to ignore such defects when only the thickness is detected at the fixed edge position of the glass through manual measurement, resulting in errors in the detected results, making it difficult to accurately control the production line, resulting in low accuracy of production line control based on photovoltaic glass thickness detection. Summary of the invention

[0004] The present application provides a production line control method, system, electronic device and storage medium based on photovoltaic glass thickness detection, which has the effect of improving the accuracy of production line control based on photovoltaic glass thickness detection.

[0005] In a first aspect, the present application provides a production line control method based on photovoltaic glass thickness detection, comprising: Obtain spectral information of the glass surface to be inspected in the target production line; Determine a plurality of detection points on the surface of the glass to be detected, and determine light characteristics and glass contour characteristics corresponding to each of the detection points based on the spectral information; Inputting the light feature and the glass profile feature into a glass thickness detection model to obtain the initial thickness of the glass to be detected at each detection point, and calculating the target thickness of the glass to be detected based on the initial thickness of each detection point; Collecting the production parameters of the glass to be tested, and calculating the production lead-out amount of the glass to be tested in combination with the production parameters and the target thickness; The glass output of the target production line is adjusted according to the difference between the production output and the standard production output.

[0006] In a second aspect of the present application, a production line control system based on photovoltaic glass thickness detection is provided, the system comprising: A spectral information acquisition module is used to acquire spectral information of the glass surface to be inspected in the target production line; A feature extraction module, used to determine a plurality of detection points on the surface of the glass to be detected, and determine the light feature and glass contour feature corresponding to each of the detection points based on the spectral information; A thickness detection module, used for inputting the light feature and the glass profile feature into a glass thickness detection model, obtaining the initial thickness of the glass to be detected corresponding to each detection point, and calculating the target thickness of the glass to be detected based on the initial thickness of each detection point; The production adjustment module is used to collect the production parameters of the glass to be tested, and calculate the production output of the glass to be tested in combination with the production parameters and the target thickness; and adjust the glass output of the target production line according to the difference between the production output and the standard production output.

[0007] In a third aspect of the present application, an electronic device is provided, comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein the program can implement a production line control method based on photovoltaic glass thickness detection when loaded and executed by the processor.

[0008] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements a production line control method based on photovoltaic glass thickness detection.

[0009] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: By adopting the above technical solutions, spectral information is obtained and light characteristics and contour characteristics are extracted. The thickness detection model is applied to perform intelligent thickness measurement on multiple points. Through model calculation, more accurate glass thickness detection is achieved, providing key parameter input for subsequent production control. Based on the initial thickness of multiple detection points, the target thickness is calculated, the overall glass thickness level is evaluated, and the error of single-point measurement is avoided. The target thickness reflects the overall quality status of the current product. The collection of production parameters and the calculation of the lead-out amount realize the effective linkage between the detection and production process, and active process control is performed based on the production input amount inferred from the target thickness. After comparing the lead-out amount with the standard value, the real-time lead-out amount of the current production line is dynamically adjusted, and a closed-loop control mechanism is established. By continuously optimizing and approaching the standard, the waste of raw materials is effectively reduced and the manufacturing cost is reduced. The images of the glass production process are captured by multiple detection points as a whole, and the glass thickness is accurately detected in combination with model calculation, so as to dynamically adjust the production conditions of the production line, thereby achieving the effect of improving the accuracy of production line control based on photovoltaic glass thickness detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 It is a flow chart of a production line control method based on photovoltaic glass thickness detection provided in an embodiment of the present application; Figure 2 It is a structural schematic diagram of a production line control system based on photovoltaic glass thickness detection provided in an embodiment of the present application; Figure 3 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application.

[0011] Description of reference numerals: 300, electronic device; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. DETAILED DESCRIPTION

[0012] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0013] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0014] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0015] The present application embodiment provides a production line control method based on photovoltaic glass thickness detection. In one embodiment, please refer to Figure 1 , Figure 1This is a flow chart of a production line control method based on photovoltaic glass thickness detection provided by an embodiment of the present application. The method can be implemented by a computer program, which can be integrated into an application or run as an independent tool application. The method can also be implemented by a single-chip microcomputer or run on a production line control system based on photovoltaic glass thickness detection based on a von Neumann system. Specifically, the method can include the following steps: Step 101: Acquire spectral information of a glass surface to be inspected in a target production line.

[0016] Among them, spectral information refers to the light intensity distribution data of the glass to be tested in the full wavelength range. In the embodiment of the present application, spectral information can be understood as a spectral curve composed of the light flux values ​​at different wavelengths obtained by measuring the spectroscopic system. It reflects the transmission and absorption of the glass sample to be tested at each wavelength. Spectral information contains the optical properties caused by changes in glass material and thickness, which can provide a data source for the subsequent establishment of light characteristics, and be used as an input for the prediction of the glass thickness detection model, thereby realizing accurate detection of glass thickness distribution.

[0017] Specifically, in order to obtain accurate and complete spectral information of the glass to be tested, a point laser light source and a spectral confocal detection system are set up during implementation. The point laser light source generates monochromatic light that is sequentially focused on each tiny area point on the surface of the glass sample. The spectral confocal system includes a spectroscopic optical element, a confocal lens group and a detector group, which can realize optical detection of the same sample point in the full wavelength range. By scanning the entire surface of the glass to be tested point by point with a point laser, the spectral confocal system can obtain the light intensity data of the sample point at all wavelengths for each scan of a sample point. Repeating this process, the spectral confocal system finally completes the point-by-point scanning of each sample point on the entire surface of the glass to be tested, and constructs a complete spectral information covering the full wavelength range. In this way, the combined application of point laser and spectral confocal technology can realize accurate optical detection of each tiny area of ​​the glass, obtain real and reliable spectral information, and provide a key optical input source and data support for the establishment of the subsequent thickness calculation model, thereby ensuring the accuracy of photovoltaic glass thickness detection.

[0018] Step 102: Determine a plurality of detection points on the glass surface to be detected, and determine the light characteristics and glass contour characteristics corresponding to each detection point based on the spectral information.

[0019] Among them, the detection point refers to a representative point predetermined on the surface of the glass to be detected. In the embodiment of the present application, it can be understood as a point with a certain distribution density obtained by segmenting the spectral information using a grid or other method, which is used to represent a local area on the glass surface. By extracting the light characteristics and contour characteristics of the detection point, the glass thickness of the corresponding area can be inferred.

[0020] Light characteristics refer to visual information that reflects the propagation effect of light on the surface of an object. In the embodiment of the present application, it can be understood as the image characteristics of reflected light and refracted light extracted from the spectral information sub-region of each detection point, which is used to represent the propagation of light in the local area of ​​​​glass corresponding to the detection point.

[0021] The glass contour feature refers to the visual feature that reflects the contour information of the glass surface shape. In the embodiment of the present application, it can be understood as the local curvature value of the glass edge calculated at each detection point position, which is used to represent the contour of the glass surface shape in the area surrounding the detection point.

[0022] Specifically, in order to comprehensively reflect the thickness distribution of the glass to be tested, it is necessary to set multiple detection points on its surface and extract light features and contour features at each point. Grid segmentation and other methods are used to determine multiple evenly distributed detection points on the spectral information to cover different areas of the glass. Then for each point, its surrounding sub-image is extracted, and the reflected light and refracted light information is determined as the light features of the point by analyzing the pixel distribution and color changes of the sub-image. At the same time, edge detection is performed to determine the glass contour, and the contour curvature of the local area of ​​each point is calculated as the contour feature. Through this process, each detection point obtains rich light features and contour features as input. These visual information can comprehensively reflect the glass thickness of each point area. Setting multi-point detection to cover the entire surface, extracting the light and shape features corresponding to the points, the thickness distribution of the glass can be inferred.

[0023] Based on the above embodiment, as an optional embodiment, in step 102: determining multiple detection points on the glass surface to be detected, and determining the light characteristics and glass contour characteristics corresponding to each detection point based on the spectral information, this step may also include the following steps: Step 201: For the glass surface to be inspected, a detection point is set at a preset interval to obtain a plurality of detection points on the glass surface to be inspected.

[0024] Specifically, in order to fully reflect the thickness distribution of the glass to be detected, it is necessary to set multiple detection points on the glass surface, which are point laser irradiation areas for detecting the glass thickness. First, measure the length of the glass to be detected, and then calculate the number of detection points that can be set on the surface of the glass to be detected according to the preset detection point interval. Then, starting from the upper left vertex of the glass surface, detection points are set in sequence along the length direction of the glass at intervals of d, and finally n uniformly distributed detection points are determined on the glass surface. The purpose of setting multiple detection points in this way is to make the detection area covered by each detection point the same size, so as to fully reflect the thickness information at different positions on the glass surface. When the spectral information of each detection point is extracted later, it can be correctly corresponded to the area where each detection point is located, and the thickness distribution of the entire glass can be comprehensively judged through the spectral information of these detection points, avoiding detection errors caused by relying only on the detection of a single point. Therefore, the setting method of the detection point can achieve the effect of more accurate detection of the thickness distribution of photovoltaic glass.

[0025] Step 202: extracting the spectral center point position and spectral shape of the area corresponding to each detection point in the spectral information; using the spectral center point position and spectral shape corresponding to each detection point as the light feature corresponding to each detection point.

[0026] Among them, the center point position of the spectrum refers to the midpoint position of the main light transmission interval in the spectrum information, and is a characteristic value reflecting the distribution center of the spectrum graph. In the embodiment of the present application, the center point position of the spectrum can be understood as the position coordinate of the center of the spectrum curve, indicating a point where the spectrum intensity distribution is relatively concentrated. This can fully reflect the average optical characteristics of each detection point area. The center point position of the spectrum is related to parameters such as glass material and thickness, and is input into the thickness detection model as a component of the light characteristics to help the model determine the initial thickness of the glass in each detection point area.

[0027] Spectral shape refers to the peak distribution form of the spectral curve, which is a characteristic value that reflects the profile of the spectral graph. In the embodiment of the present application, the spectral shape can be understood as the shape profile composed of the inflection point, peak value, valley value, etc. of the spectral curve, which indicates the change of spectral intensity with wavelength. This can reflect the specific transmission distribution of the spectral curve in each detection point area. The spectral shape is related to parameters such as the internal structure of the glass and thickness change. It is input into the thickness detection model as a component of the light characteristics to help the model determine the initial thickness of the glass in each detection point area.

[0028] Specifically, the spectral information reflects the light transmission and absorption characteristics of the glass to be tested at different wavelengths. Among them, the center point position of the spectrum represents the main light transmission interval, and the spectrum shape represents the peak distribution of the light transmission curve. These two features can effectively reflect the material characteristics of the glass. Therefore, it is necessary to extract these two features of each detection point area as light features. First, according to the spectral information and the preset interval range, the center position of the spectrum curve corresponding to each detection point is calculated, that is, the center point position of the spectrum. Then, the shape of the spectrum curve of each detection point area is analyzed, such as the number of peaks, half-height width, etc., to obtain the characteristic value representing the shape of the spectrum. These two parts of the features constitute the light features of each detection point. The center point position and the spectrum shape of the spectrum are extracted as light features because they can fully reflect the light transmission characteristics of each detection point area, which is related to the material characteristics of the glass, so that they can be passed as input to the glass thickness detection model. The model can judge the initial thickness of each detection point based on these light features. If only a single feature is extracted, it is difficult to accurately judge the thickness. Therefore, it is necessary to combine two spectral features as light features to improve the accuracy of subsequent thickness detection.

[0029] Step 203: Perform edge detection on the spectral information to obtain a contour image of the glass to be detected; based on the contour image, determine the contour curvature corresponding to each detection point as the glass contour feature.

[0030] Among them, edge detection refers to the image processing technology for identifying and locating edge lines in an image. In the embodiment of the present application, it can be understood as using algorithm tools to extract the edge contour of the glass piece from the spectral information, which is used to determine the outer shape boundary of the glass and is a prerequisite for obtaining contour features.

[0031] The contour image refers to a binary image containing the target edge contour information. In the embodiment of the present application, it can be understood as an image containing only the outer contour of the glass sheet obtained after processing the spectral information through the edge detection algorithm. It is used to represent the boundary shape of the glass and is the basis for calculating the contour curvature.

[0032] The contour curvature refers to the curvature size representing the shape of the edge contour. In the embodiment of the present application, it can be understood as calculating the curvature value of the local area of ​​the edge where each detection point is located on the contour image, which is used to reflect the shape characteristics of the contour around the detection point as the contour feature of the detection point.

[0033] Specifically, in order to obtain the glass contour features of each detection point, it is necessary to first detect the edge of the spectral information, extract the glass contour line, and then calculate the local contour curvature information of each point. The spectral information is edge detected using algorithms such as Canny to obtain the edge contour image of the glass. Then, based on the contour image, the curvature of the local area contour of each detection point is measured as the contour feature through the curvature calculation algorithm. This process analyzes the surface features from the perspective of geometric shape, and the contour curvature reflects the shape changes of the surface at different positions. Combined with light features, it can provide richer visual input. The contour feature provides shape information for each point, which is conducive to improving the diversification and accuracy of thickness detection. This step obtains unique contour feature information for each detection point through image edge analysis and local curvature calculation, which complements the light features and provides rich visual input for thickness detection to improve the accuracy of the final result.

[0034] Step 103: Input the light characteristics and the glass contour characteristics into the glass thickness detection model to obtain the initial thickness of the glass to be detected at each detection point, and calculate the target thickness of the glass to be detected based on the initial thickness of each detection point.

[0035] Among them, the glass thickness detection model refers to a glass thickness prediction model based on machine learning. In the embodiment of the present application, it can be understood as a machine learning model represented by a neural network. The input is light features and contour features, and the output is the thickness prediction result of each detection point. It is used to realize the mapping from the visual features of spectral information to glass thickness. It is the core model of optical thickness detection.

[0036] The initial thickness refers to the thickness value of each detection point obtained by the thickness detection model based on the visual characteristics of the spectral information. In the embodiment of the present application, it can be understood as the predicted result of the thickness of each detection point output by the model, which is used to represent the intermediate result of the thickness detection of each point area. The final target thickness can be calculated by synthesizing the initial thickness of each detection point.

[0037] The target thickness refers to the overall thickness value of the glass obtained by statistically calculating the initial thickness of all detection points. In the embodiment of the present application, it can be understood as the final thickness result of the glass calculated based on the initial thickness of each detection point using statistical methods such as the mean method. It is used to represent the average thickness level of the entire glass to be detected, and is the final output result of optical thickness detection.

[0038] Specifically, in order to calculate the thickness distribution of the glass, it is necessary to input the light features and contour features into the thickness detection model to obtain the initial thickness of each detection point, and further calculate the final target thickness. The light features and contour features extracted from each detection point are combined and input into the neural network model for glass thickness detection. The model predicts the initial thickness value of each detection point based on rich visual features. Then, based on the initial thickness of all detection points, statistical methods such as the mean method and the median method are applied to comprehensively calculate the final target thickness of the glass. This process makes full use of the rich visual features, and the setting of the detection points also improves the sampling coverage. The model predicts the initial thickness distribution, and then calculates the target thickness, realizing accurate measurement of the overall glass thickness.

[0039] Based on the above embodiment, as an optional embodiment, in step 103: inputting the light characteristics and the glass contour characteristics into the glass thickness detection model, before this step, the following steps may also be included: Step 301: Acquire multiple historical glass samples and the glass thickness corresponding to each historical glass sample; process the spectral information of each historical glass sample to obtain a light feature sample and a glass profile feature sample of each historical glass sample.

[0040] Among them, historical glass samples refer to real glass sheet samples used for model training. In the embodiment of the present application, they can be understood as a set of glass samples with various parameter specifications collected, which are used to provide the model with sufficient training data. By marking their actual thickness, the model can learn the relationship between glass thickness and visual features.

[0041] Specifically, in order to train an accurate thickness detection model, it is necessary to prepare a large number of high-quality historical glass samples and corresponding thickness labels, and extract features from the samples in advance to improve training efficiency. First, real glass of various specifications is collected as historical samples, and the thickness values ​​of the samples are accurately measured using a contact thickness gauge to establish a sample data set. Then, for the spectral information of each sample, the light features and contour features are extracted to form a feature sample set. This step provides key support for training a high-performance thickness detection model through quality-assured sample collection and feature extraction, enabling the model to learn accurate thickness mapping relationships.

[0042] Step 302: taking the light feature samples, glass contour feature samples and glass thickness of each historical glass sample as training data, and combining the training data to perform model training on a preset neural network model to obtain a glass thickness detection model after model training.

[0043] Among them, the preset neural network model refers to the neural network structure framework pre-set before model training. In the embodiment of the present application, it can be understood as an initialized neural network model structure, which is used to gradually learn and optimize parameters during the training process, and finally form a neural network model for glass thickness prediction.

[0044] Specifically, in order to obtain a prediction model for glass thickness detection, it is necessary to use historical samples for supervised model training to establish the correspondence between features and thickness values. Prepare a preset neural network model structure. Then extract the light features, contour features and corresponding thickness values ​​of each historical glass sample as supervised training data, and input them into the preset model for training. Through the learning of a large number of samples, the model gradually optimizes the parameters and establishes a mapping relationship between features and thickness. After iterative training, a refined glass thickness detection model is finally obtained. This process uses a supervised learning method and rich historical sample training to enable the model to master the inherent correspondence rules between image features and true thickness values. The model can efficiently predict the thickness of new samples, realize the intelligent mapping from spectral information to thickness, and obtain a high-precision thickness prediction model through targeted model training, which provides data support for optical thickness detection.

[0045] Based on the above embodiment, as an optional embodiment, in step 103: based on the initial thickness of each detection point, the target thickness of the glass to be detected is calculated. This step may also include the following steps: Step 303: Calculate the average thickness of the initial thickness of each detection point, and determine the thickness difference between the initial thickness of each detection point and the average thickness.

[0046] Specifically, in order to evaluate the performance of the thickness detection model, it is necessary to calculate the average value of the initial thickness of the test results and determine the difference between each test point and the average value, so as to evaluate the comprehensive accuracy of the model detection. First, based on the initial thickness of each test point output by the model, the average value of the thickness of all test points is calculated as the average level of the initial thickness. Then, the difference between the initial thickness of each test point and the average value is calculated to indicate the deviation of the thickness detection at each point. This process evaluates the model's ability to detect thickness at different test points. Through the analysis of the average deviation, the overall detection stability and local refined detection capabilities of the model can be judged. This step comprehensively evaluates the performance of the thickness detection model based on the statistical analysis of the initial thickness, provides quantitative analysis support for model optimization, and improves the comprehensive detection accuracy of the model.

[0047] Step 304: among the detection points, screen out the detection points whose thickness difference is greater than or equal to the thickness difference threshold; perform weighted summation on the initial thickness of the detection points after screening out to obtain the target thickness of the glass to be detected.

[0048] Specifically, in order to filter out points with large detection errors and improve the accuracy of target thickness calculation, it is necessary to screen out detection points with too large initial thickness differences and only use more accurate detection points to calculate the target thickness. First, a thickness difference threshold is preset. Then, among all detection points, the detection points whose initial thickness differs from the average thickness by more than the threshold are screened out. Only detection points with small differences are retained, and their initial thicknesses are weighted summed to calculate the final target thickness. This process filters out outliers with large detection errors by setting an error tolerance, thus avoiding their interference with the final result. Only using points with higher precision to calculate the target thickness improves the overall accuracy of the detection. This step is based on error filtering and key sampling to extract high-quality detection points, accurately calculate the final target thickness, and achieve precision in the detection results.

[0049] Exemplarily, the target thickness of the glass to be inspected is also measured by moving average. The point laser moving average method is used to calculate the target thickness because the point laser can perform local precise measurement and can achieve comprehensive detection by scanning different positions, and the moving average can effectively filter the random errors in the detection values ​​of each single point and improve the detection accuracy. First, multiple detection points are set on the glass surface, and the initial thickness value of the corresponding position is measured by using a point laser measurement sensor at each detection point. Then, the point laser is moved to each detection point in turn along a certain path for measurement. At the same time, a thickness sampling time window is set, and the thickness values ​​of each detection point obtained within the time window are averaged to obtain the current average thickness. After that, the sampling window continues to move to the next position, and the averaging process is repeated to obtain the next average thickness. By continuously moving the sampling window, the average thickness value of the illumination point moving on each detection point can be obtained. Compared with single point measurement, the moving average can effectively offset the influence of individual errors of each point and calculate a more accurate and smooth target thickness. Compared with directly averaging all detection points, the moving scanning of the point laser also more comprehensively reflects the thickness conditions at different positions. This method based on point laser moving average not only has the accuracy of point measurement, but also can improve the anti-interference ability of global detection through moving average, so as to calculate more accurate and reliable target thickness and provide precise control parameters for subsequent production line control.

[0050] Based on the above embodiment, as an optional embodiment, in step 304: weighted summing of the initial thicknesses of each detection point after screening is performed to obtain the target thickness of the glass to be detected, this step may also include the following steps: Step 314: determine the center point of the glass to be inspected; and determine the weight coefficient of each inspection point after screening out according to the distance between each remaining inspection point and the center point.

[0051] Specifically, in order to make the target thickness calculation more accurate and reliable, it is necessary to set weights according to the distance between the detection point and the center point, and give the points near the center a greater calculation weight. Determine the geometric center point of the glass. Then calculate the distance between each remaining detection point and the center point. The farther the distance, the smaller the weight coefficient. Finally, according to the weight coefficient of each point, the initial thickness is weighted and summed to obtain an objective and accurate target thickness. The weight distribution scheme of the detection points is designed in this process. Since the light effect in the center point area is smaller and the detection is more accurate, setting a larger weight can improve the reliability of the target thickness. Through the idea of ​​weighting the center point, the importance division of the detection points is scientifically designed, so that the key area plays a leading role, the overall thickness is accurately calculated, and the detection accuracy is improved.

[0052] Step 324: Based on the weight coefficients of the detected points after screening, weighted sum is performed on the initial thicknesses to obtain the target thickness of the glass to be detected.

[0053] Specifically, in order to utilize the prediction results of each detection point, it is necessary to perform a weighted combination of the initial thicknesses according to the set weight coefficients to calculate the final target thickness. The weight of each point is determined based on the distance between the detection point and the center point. Then, the initial thickness of each detection point after screening is calculated by proportional weighted summation, that is, the initial thickness of each detection point is multiplied by its weight coefficient to calculate the sum of the weighted thicknesses of all detection points. Finally, the target thickness of the glass to be tested is obtained. This process designs the division of the contribution of each detection point so that more accurate points play a leading role, effectively improving the calculation accuracy of the target thickness. Through the weighted summation algorithm, the prediction of each detection point is scientifically utilized to obtain an accurate and reliable final target thickness, further improving the overall effect of optical thickness detection.

[0054] Step 104: Collect the production parameters of the glass to be tested, and calculate the production lead-out amount of the glass to be tested in combination with the production parameters and the target thickness.

[0055] Among them, production parameters refer to the key technical indicators in the glass manufacturing process. In the embodiments of the present application, they can be understood as parameters such as glass formula, melting temperature, calender pressure, flow rate, etc., which are used to reflect the various settings of the current production process. These parameters will directly affect the thickness of the final glass.

[0056] The production lead-out amount refers to the optimized outflow amount of glass liquid in the glass manufacturing process calculated based on the test results. In the embodiment of the present application, it can be understood as a reasonable amount of molten glass calculated by reverse calculation based on the target thickness, which is used to adjust the glass flow rate during the liquid injection process to make the thickness distribution of the liquid glass more uniform and reasonable, so as to further optimize the entire production process.

[0057] Specifically, in order to achieve closed-loop control of detection and production, it is necessary to collect production parameters, calculate the optimized production lead-out amount based on the parameters and detection results, and adjust the production process in real time. Collect various key parameters from the production execution system, such as formula ratio, melting temperature, calender pressure, etc. Then input the target thickness data obtained by the detection, combine the production parameters, establish a metering model, and calculate the corrected reasonable lead-out amount according to the proportional relationship. Finally, the injection process is controlled according to this lead-out amount to achieve closed-loop optimization of the production flow. This process makes full use of detection feedback and realizes active control from results to processes. The accurately calculated lead-out amount adjusts and optimizes the production process, which can reduce the waste of materials required for glass production and reduce costs.

[0058] Based on the above embodiment, as an optional embodiment, in step 104: combining the production parameters and the target thickness to calculate the production lead-out amount of the glass to be tested, this step may also include the following steps: Step 401: Determine the cross-sectional area of ​​the glass to be inspected based on the width of the glass to be inspected and the target thickness.

[0059] Specifically, in order to calculate the yield of glass, it is necessary to first determine the cross-sectional area of ​​the glass based on the detection width and target thickness. First, the width data of the glass to be detected is analyzed and extracted from the image. Then, combined with the target thickness obtained by the detection, the cross-sectional area of ​​the glass to be detected is calculated using the cross-sectional area = width × thickness. This process utilizes the width and thickness results of the detection, avoids repeated detection, and directly calculates the cross-sectional area based on the image analysis parameters. The cross-sectional area reflects the quality specifications of the current unit of glass. The calculation is simplified through the comprehensive use of parameters, and the cross-sectional area of ​​the glass is accurately and efficiently obtained using the detection image, laying the foundation for subsequent yield calculations.

[0060] Step 402: multiply the cross-sectional area of ​​the glass to be inspected by the moving speed of the glass to be inspected on the production line to obtain the lead-out volume of the glass to be inspected within a preset unit time.

[0061] Among them, the moving speed refers to the actual running speed of the glass on the conveyor belt. In the embodiment of the present application, it can be understood as the linear speed of the glass to be tested on the production line, which is used to calculate the output volume per unit time and evaluate the production efficiency of the production process.

[0062] Specifically, in order to further calculate the output of glass, it is necessary to calculate the output volume per unit time based on the cross-sectional area and the moving speed. First, detect and obtain the moving speed of the glass on the conveyor belt. Then multiply the cross-sectional area obtained in the previous step by the speed, and apply the formula: volume = cross-sectional area × speed × time to calculate the output volume of the glass in the preset unit time. Through this process, the speed parameter is utilized and the cross-sectional area is expanded to a three-dimensional output calculation. The output volume per unit time intuitively reflects the output level under the current production conditions. This step accurately and efficiently calculates the output of glass per unit time through comprehensive consideration of the moving speed, providing a data reference for capacity evaluation and process control.

[0063] Step 403: Obtain the glass density of the glass to be tested; and use the product of the lead-out volume and the glass density as the production lead-out amount of the glass to be tested.

[0064] Among them, glass density refers to the inherent physical parameters of glass of different materials. In the embodiment of the present application, it can be understood as the density value of the glass material currently used in production, which is used to convert the calculated volume output into mass output and evaluate the total production consumption per unit time.

[0065] Specifically, in order to achieve accurate conversion from volume to mass, it is necessary to obtain material density parameters and perform dimensional conversion to obtain accurate mass extraction. First, find the density value of the glass type currently used in production from the material database. Then take the volume output per unit time calculated in the previous step, multiply it by the glass density, and calculate the total mass extraction per unit time through dimensional conversion. This process utilizes material density parameters to achieve accurate dimensional conversion and obtains a mass extraction result that is consistent with the actual consumption. This result intuitively reflects the total consumption of glass raw materials required for production per unit time. This step realizes accurate calculation from volume to mass based on material properties, laying the foundation for subsequent production measurement and process control.

[0066] Step 105: According to the difference between the production lead-out amount and the standard production lead-out amount, adjust the glass lead-out amount of the target production line.

[0067] Specifically, in order to realize the production process control of detection feedback, it is necessary to adjust the glass lead-out amount of the production line in real time according to the difference between the calculated lead-out amount and the standard lead-out amount. First, the standard lead-out amount parameters of the production line are preset. Then the difference between the lead-out amount obtained in the current detection link and the standard amount is calculated. If the detected lead-out amount is greater than the standard amount, the actual lead-out amount of the production line is reduced, otherwise the lead-out amount is increased. In this way, the actual production volume can gradually approach the optimized standard lead-out amount. This process builds a closed-loop control between detection and production. The real-time adjustment of the lead-out amount optimizes the production process and effectively improves the stability of product quality. This step realizes active control based on the feedback of the lead-out amount, so that the detection results can accurately guide the production process.

[0068] Based on the above embodiment, as an optional embodiment, in step 105: adjusting the glass lead-out amount of the target production line according to the difference between the production lead-out amount and the standard production lead-out amount, this step may also include the following steps: Step 501: According to the difference, the corresponding basic glass lead-out amount is matched in a preset lead-out amount mapping table.

[0069] Among them, the preset lead-out quantity mapping table refers to a database that stores the relationship between the lead-out quantity difference and the corresponding basic lead-out quantity. In the embodiment of the present application, it can be understood as a query table containing multiple groups of lead-out quantity differences and their matching benchmark lead-out quantities, which is used to quickly find and determine the current benchmark lead-out quantity through the difference, simplify the calculation process, and improve the efficiency of parameter acquisition.

[0070] The basic glass output refers to the most basic glass production flow obtained by querying the difference of the current detection results. In the embodiment of the present application, it can be understood as the current optimal initial output value obtained by matching the mapping table, which is used as the basic input for subsequent output optimization calculations. By fine-tuning on this basis, the optimal production output can be further fitted.

[0071] Specifically, in order to quickly obtain a preliminary and reasonable lead-out amount, it is necessary to directly search for the matching basic lead-out amount based on the constructed mapping table. First, a large amount of historical data is collected to establish a corresponding relationship table between the lead-out amount difference and the basic lead-out amount. Then, the lead-out amount difference calculated currently is taken, and its corresponding basic lead-out amount value is searched in the mapping table. Through the pre-established table lookup correspondence, a fast conversion is achieved, which avoids repeated modeling and simplifies the determination of the basic lead-out amount. The obtained basic amount lays a data foundation for subsequent precise calculations.

[0072] Step 502: If the production output is greater than or equal to the standard production output, the base glass output is subtracted from the current glass output of the target production line.

[0073] Specifically, when the calculated production lead-out amount is greater than or equal to the standard lead-out amount, the glass lead-out amount of the current target production line needs to be reduced accordingly, specifically by subtracting the obtained basic lead-out amount from the current lead-out amount. Determine whether the production lead-out amount is greater than or equal to the standard lead-out amount. If so, take the real-time lead-out amount of the current production line and subtract the basic lead-out amount value obtained in the previous step to obtain the corrected new lead-out amount result, thereby realizing the automatic reduction and optimization of the excess lead-out amount. The reduction in the lead-out amount can effectively reduce raw material consumption and improve production efficiency. The new lead-out amount is more in line with quality requirements and is conducive to improving product qualification rate. This step realizes the automatic adjustment of the excess lead-out amount based on threshold judgment, and realizes the optimization of production lead-out by utilizing the adjustment range of the basic lead-out amount, thereby reducing production costs.

[0074] Step 503: If the production output is less than the standard production output, the current glass output of the target production line is increased by the basic glass output.

[0075] Specifically, when the calculated production lead-out amount is less than the standard lead-out amount, it is necessary to increase the glass lead-out amount of the current target production line accordingly, specifically, to increase the current lead-out amount by the obtained basic lead-out amount. First, determine whether the production lead-out amount is less than the standard lead-out amount. If so, take the real-time lead-out amount of the current production line, and add the basic lead-out amount value obtained in the previous step to obtain the corrected new lead-out amount result. Since a small lead-out amount will lead to unqualified glass production quality when the production speed is stable, the process realizes the automatic increase and optimization of the gap lead-out amount by increasing the lead-out amount. The increase in the lead-out amount can control the quality of glass production during the glass production process. The new lead-out amount is more in line with the quality standards and is conducive to improving the product qualification rate. This step realizes the automatic adjustment of the gap lead-out amount based on the threshold judgment, and realizes the production lead-out optimization by utilizing the adjustment range of the basic lead-out amount.

[0076] Reference Figure 2 , is a production line control system based on photovoltaic glass thickness detection provided by an embodiment of the present application, the system includes: a spectral information acquisition module, a feature extraction module, a thickness detection module, and a production adjustment module, wherein: A spectral information acquisition module is used to acquire spectral information of the glass surface to be inspected in the target production line; A feature extraction module is used to determine multiple detection points on the glass surface to be detected, and based on the spectral information, determine the light characteristics and glass contour characteristics corresponding to each detection point; The thickness detection module is used to input the light characteristics and the glass profile characteristics into the glass thickness detection model to obtain the initial thickness of the glass to be detected at each detection point, and calculate the target thickness of the glass to be detected based on the initial thickness of each detection point; The production adjustment module is used to collect the production parameters of the glass to be tested, and calculate the production output of the glass to be tested in combination with the production parameters and the target thickness; according to the difference between the production output and the standard production output, the glass output of the target production line is adjusted.

[0077] On the basis of the above embodiments, the feature extraction module is also used to set a detection point at a preset distance for the glass surface to be detected, so as to obtain multiple detection points on the glass surface to be detected; extract the spectral center point position and spectral shape of the area corresponding to each detection point in the spectral information; use the spectral center point position and spectral shape corresponding to each detection point as the light feature corresponding to each detection point; perform edge detection on the spectral image in the spectral information to obtain the contour image of the glass to be detected; based on the contour image, determine the contour curvature corresponding to each detection point as the glass contour feature.

[0078] On the basis of the above-mentioned embodiment, the thickness detection module is also used to obtain multiple historical glass samples and the glass thickness corresponding to each historical glass sample; process the spectral information of each historical glass sample to obtain the light feature sample and the glass contour feature sample of each historical glass sample; use the light feature sample, glass contour feature sample and glass thickness of each historical glass sample as training data, and perform model training on a preset neural network model in combination with the training data to obtain a glass thickness detection model after model training.

[0079] Based on the above embodiment, the thickness detection module is also used to calculate the average thickness of the initial thickness of each detection point, and determine the thickness difference between the initial thickness of each detection point and the average thickness; among each detection point, the detection points with a thickness difference greater than or equal to the thickness difference threshold are screened out; the initial thickness of each detection point after screening is weightedly summed to obtain the target thickness of the glass to be detected.

[0080] Based on the above embodiment, the thickness detection module is also used to determine the center point of the glass to be detected; determine the weight coefficient of each detection point after screening out according to the distance between each remaining detection point and the center point; and perform weighted summation on each initial thickness based on the weight coefficient of each detection point after screening out to obtain the target thickness of the glass to be detected.

[0081] On the basis of the above-mentioned embodiment, the production adjustment module is also used to determine the cross-sectional area of ​​the glass to be inspected based on the width of the glass to be inspected and the target thickness; multiply the cross-sectional area of ​​the glass to be inspected by the moving speed of the glass to be inspected on the production line to obtain the lead-out volume of the glass to be inspected within a preset unit time; obtain the glass density of the glass to be inspected; and use the product of the lead-out volume and the glass density as the production lead-out amount of the glass to be inspected.

[0082] Based on the above embodiment, the production adjustment module is also used to match the corresponding basic glass output in the preset output mapping table according to the difference; if the production output is greater than or equal to the standard production output, the basic glass output is subtracted from the current glass output of the target production line; if the production output is less than the standard production output, the basic glass output is increased by the current glass output of the target production line.

[0083] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0084] The present application also discloses an electronic device. Figure 3 , Figure 3 The electronic device 300 may include: at least one processor 301 , at least one network interface 304 , a user interface 303 , a memory 305 , and at least one communication bus 302 .

[0085] The communication bus 302 is used to realize the connection and communication between these components.

[0086] The user interface 303 may include a display interface and a camera interface. Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0087] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0088] Among them, the processor 301 may include one or more processing cores. The processor 301 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Optionally, the processor 301 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 301 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface diagrams and applications, etc.; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 301, and it can be implemented separately through a chip.

[0089] Among them, the memory 305 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may optionally also be at least one storage device located away from the aforementioned processor 301. Refer to Figure 3 , the memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a production line control method based on photovoltaic glass thickness detection.

[0090] exist Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 301 can be used to call the application program stored in the memory 305 for a production line control method based on photovoltaic glass thickness detection. When executed by one or more processors 301, the electronic device 300 executes one or more methods in the above-mentioned embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited to the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0091] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0092] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0093] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0094] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0095] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0096] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0097] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not recorded in the present disclosure. The description and examples are to be regarded as exemplary only.

Claims

1. A production line control method based on photovoltaic glass thickness detection, characterized in that: include: Obtain spectral information of the glass surface to be inspected in the target production line; Determine a plurality of detection points on the surface of the glass to be detected, and determine light characteristics and glass contour characteristics corresponding to each of the detection points based on the spectral information; Inputting the light feature and the glass profile feature into a glass thickness detection model to obtain the initial thickness of the glass to be detected at each detection point, and calculating the target thickness of the glass to be detected based on the initial thickness of each detection point; Collecting the production parameters of the glass to be tested, and calculating the production lead-out amount of the glass to be tested in combination with the production parameters and the target thickness; The glass output of the target production line is adjusted according to the difference between the production output and the standard production output.

2. The production line control method based on photovoltaic glass thickness detection according to claim 1 is characterized in that: The step of determining a plurality of detection points on the surface of the glass to be detected, and determining light features and glass contour features corresponding to each of the detection points based on the spectral information, includes: For the glass surface to be inspected, a detection point is set at a preset distance to obtain a plurality of detection points on the glass surface to be inspected; Extracting the spectral center point position and spectral shape of the area corresponding to each detection point in the spectral information; The center point position and the shape of the spectrum corresponding to each detection point are used as the light feature corresponding to each detection point; Performing edge detection on the spectral image in the spectral information to obtain a contour image of the glass to be detected; Based on the contour image, the contour curvature corresponding to each of the detection points is determined as the glass contour feature.

3. The production line control method based on photovoltaic glass thickness detection according to claim 1 is characterized in that: Before inputting the light feature and the glass contour feature into the glass thickness detection model, the method further includes: Acquire a plurality of historical glass samples and the glass thickness corresponding to each of the historical glass samples; Processing the spectral information of each of the historical glass samples to obtain a light feature sample and a glass profile feature sample of each of the historical glass samples; The light feature samples, glass contour feature samples and the glass thickness of each of the historical glass samples are used as training data, and the preset neural network model is trained in combination with each of the training data to obtain the glass thickness detection model after model training.

4. The production line control method based on photovoltaic glass thickness detection according to claim 1 is characterized in that: The step of calculating the target thickness of the glass to be inspected based on the initial thickness of each inspection point comprises: Calculating the average thickness of the initial thicknesses of the detection points, and determining the thickness difference between the initial thickness of the detection points and the average thickness; Among the detection points, screening out the detection points whose thickness difference is greater than or equal to the thickness difference threshold; The initial thickness of each of the detection points after screening is weightedly summed to obtain the target thickness of the glass to be detected.

5. The production line control method based on photovoltaic glass thickness detection according to claim 4 is characterized in that: The step of performing weighted summation on the initial thicknesses of the removed detection points to obtain the target thickness of the glass to be detected comprises: Determine the center point of the glass to be inspected; Determine the weight coefficient of each of the detected points after screening out according to the distance between each of the remaining detected points and the central point; Based on the weight coefficient of each of the detected points after screening, each of the initial thicknesses is weightedly summed to obtain the target thickness of the glass to be detected.

6. The production line control method based on photovoltaic glass thickness detection according to claim 1 is characterized in that: The production parameters include the width of the glass to be detected and the moving speed of the glass to be detected on the production line. The production output amount of the glass to be detected is calculated by combining the production parameters and the target thickness, including: Determining the cross-sectional area of ​​the glass to be inspected based on the width of the glass to be inspected and the target thickness; The cross-sectional area of ​​the glass to be inspected is multiplied by the moving speed of the glass to be inspected on the production line to obtain the lead-out volume of the glass to be inspected in a preset unit time; Acquiring the glass density of the glass to be tested; The product of the lead-out volume and the glass density is taken as the production lead-out amount of the glass to be tested.

7. The production line control method based on photovoltaic glass thickness detection according to claim 1 is characterized in that: The step of adjusting the glass lead-out amount of the target production line according to the difference between the production lead-out amount and the standard production lead-out amount includes: According to the difference, the corresponding basic glass lead-out amount is matched in a preset lead-out amount mapping table; If the production output is greater than or equal to the standard production output, subtract the basic glass output from the current glass output of the target production line; If the production output is less than the standard production output, the current glass output of the target production line is increased by the basic glass output.

8. A production line control system based on photovoltaic glass thickness detection, characterized in that: The system comprises: A spectral information acquisition module is used to acquire spectral information of the glass surface to be inspected in the target production line; A feature extraction module, used to determine a plurality of detection points on the surface of the glass to be detected, and determine the light feature and glass contour feature corresponding to each of the detection points based on the spectral information; A thickness detection module, used for inputting the light feature and the glass profile feature into a glass thickness detection model, obtaining the initial thickness of the glass to be detected corresponding to each detection point, and calculating the target thickness of the glass to be detected based on the initial thickness of each detection point; The production adjustment module is used to collect the production parameters of the glass to be tested, and calculate the production output of the glass to be tested in combination with the production parameters and the target thickness; and adjust the glass output of the target production line according to the difference between the production output and the standard production output.

9. An electronic device, characterized in that: It includes a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the production line control method based on photovoltaic glass thickness detection as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the production line control method based on photovoltaic glass thickness detection as described in any one of claims 1 to 7 is executed.

Images