A method and system for high-precision scribing of perovskite by primary and auxiliary laser cooperation

Abstract

This application provides a method and system for high-precision scribing of perovskite using a master-auxiliary laser synergy. The method includes: using a master laser system to initially scribble a pre-defined fixed trajectory for a battery module to be cut, obtaining a preliminary scribing area; using an optical detection system to perform real-time tracking detection of the preliminary scribing area, obtaining detection data; and based on the detection data, determining a depth difference map and a width difference map of the preliminary scribing; generating a trajectory compensation map based on the depth difference map and the width difference map using a pre-defined neural network; wherein each pixel in the trajectory compensation map contains three types of compensation components: laser power, moving speed, and spot size; and correcting the auxiliary laser parameters of the auxiliary laser system based on the three types of compensation components in the trajectory compensation map to obtain a precise scribing area. This solution improves scribing accuracy and solves the problem of traditional accuracy control issues.

Description

Technical Field

[0001] This application relates to the field of laser scribing technology, specifically to a method and system for high-precision scribing of perovskites using a combination of primary and secondary lasers. Background Technology

[0002] Perovskite solar cells have become an important development direction in the new energy field due to their advantages of high photoelectric conversion efficiency and low manufacturing cost. Laser scribing (P1-P2-P3 process) is the core link in the large-scale manufacturing of its modules. It is necessary to use laser to accurately divide the conductive layer, perovskite composite layer and top electrode layer to form series sub-cell units. The scribing quality directly determines the final efficiency and long-term reliability of the module.

[0003] However, current traditional laser scribing technology is ill-suited to the unique characteristics of perovskite processing. On the one hand, perovskite films have poor thermal stability and low mechanical strength, making them extremely sensitive to laser energy; even slight energy inappropriateness can lead to film damage. On the other hand, the large-format substrates commonly used in industrial applications are prone to warping or localized unevenness due to uneven support and temperature changes, further exacerbating the risk of scribing deviations. Furthermore, the traditional process employs a "main laser processing + offline inspection" model, which not only suffers from delayed inspection and inability to correct errors in a timely manner, but also uses a crude compensation strategy, easily leading to new processing defects and severely hindering the improvement of perovskite module yield and the progress of large-scale mass production. In addition, the crude compensation strategy results in insufficient scribing accuracy. Traditional compensation often uses "uniform parameter adjustment" (such as changing the overall laser energy) without developing adaptive solutions for different error types in different areas (such as depth deviation or insufficient width). Especially when facing localized deformation of large-format substrates, it easily leads to "correction in one place, damage in multiple places," failing to meet the high-precision scribing requirements of perovskite modules. Summary of the Invention

[0004] This application aims to provide a method and system for high-precision perovskite scribing using a master-slave laser synergy, which can improve scribing accuracy and solve the problem of traditional accuracy loss of control.

[0005] The technical solution of this application is implemented as follows:

[0006] In a first aspect, embodiments of this application provide a method for high-precision scribing of perovskites using a master-slave laser synergy, the method comprising:

[0007] The main laser system performs preliminary scribing of the battery module to be cut according to a preset fixed trajectory, thus obtaining the preliminary scribing area.

[0008] The preliminary marking area is monitored in real time using an optical inspection system to obtain inspection data; and based on the inspection data, the depth difference map and width difference map of the preliminary marking are determined.

[0009] Based on the depth difference map and the width difference map, a trajectory compensation map is generated through a preset neural network; wherein, each pixel in the trajectory compensation map contains three types of compensation components: laser power, moving speed, and spot size;

[0010] Based on the three types of compensation components in the trajectory compensation diagram, the auxiliary laser parameters of the auxiliary laser system are corrected to obtain a precise line drawing area.

[0011] In the above scheme, the step of performing real-time tracking detection on the preliminary marking area using an optical detection system to obtain detection data, and determining the depth difference map and width difference map of the preliminary marking based on the detection data, includes:

[0012] The optical detection system continuously captures images of the initially marked area to obtain the original image of the initially marked area.

[0013] The original image is subjected to image binarization, 3D visual reconstruction, and high-frequency component extraction to obtain the detection data; wherein, the detection data includes actual line width data and actual depth data;

[0014] Based on the actual depth data and the actual line width data, a pixel-by-pixel comparison is performed with the preset target depth and the preset target line width, and the difference is calculated based on the pixel comparison results to obtain the depth difference map and the width difference map.

[0015] In the above scheme, the detection data also includes a burr distribution map;

[0016] The detection data is obtained by performing image binarization, 3D visual reconstruction, and high-frequency component extraction on the original image, including:

[0017] The original image is binarized to obtain a binarized image; and edge extraction and feature analysis are performed on the binarized image to obtain the actual line width data.

[0018] The original image is reconstructed using 3D vision processing technology to obtain a depth image; and pixel-level depth values ​​are calculated from the depth image to determine the actual depth data.

[0019] By using frequency domain filtering, high-frequency components are extracted and features are identified from the original image to obtain the spiky distribution map.

[0020] In the above scheme, the step of extracting high-frequency components and identifying features from the original image through frequency domain filtering to obtain the spur distribution map includes:

[0021] The high-frequency components are extracted from the original image by frequency domain filtering.

[0022] By using multiple pre-set thresholds, the high-frequency components are subjected to threshold segmentation to obtain a high-frequency truncated image;

[0023] Edge burr feature recognition is performed on the high-frequency truncated image to obtain the burr distribution map.

[0024] In the above scheme, the preset neural network includes a trajectory sub-neural network, a light intensity sub-neural network, and a correction sub-neural network;

[0025] The step of generating a trajectory compensation map based on the depth difference map and the width difference map using a preset neural network includes:

[0026] The trajectory sub-neural network is used to extract features from the width difference map and the spur distribution map to obtain a first moving speed feature map and a first spot size feature map.

[0027] The optical intensity sub-neural network is used to extract features from the depth difference map, the width difference map, and the burr distribution map to obtain a second moving speed feature map, a second spot size feature map, and a laser power feature map.

[0028] The corrective sub-neural network is used to perform feature fusion and parameter regression on the first moving speed feature map, the first spot size feature map, the second moving speed feature map, the second spot size feature map, and the laser power feature map to obtain the laser power, the moving speed, and the spot size.

[0029] The trajectory compensation map is determined based on the laser power, the moving speed, and the spot size.

[0030] In the above scheme, the step of performing feature fusion and parameter regression on the first moving speed feature map, the first spot size feature map, the second moving speed feature map, the second spot size feature map, and the laser power feature map through the corrective sub-neural network to obtain the laser power, the moving speed, and the spot size includes:

[0031] The first and second movement speed feature maps are fused using the corrective sub-neural network to obtain the final movement speed feature map.

[0032] The first spot size feature map and the second spot size feature map are fused using the corrective sub-neural network to obtain the final spot size feature map.

[0033] The final moving speed feature map, the final spot size feature map, and the laser power feature map are spliced ​​together to obtain pixel fusion features;

[0034] By using a fully connected layer and a linear function, parameter regression is performed on the pixel fusion features to obtain the laser power, the moving speed, and the spot size.

[0035] In the above scheme, the step of correcting the auxiliary laser parameters of the auxiliary laser system based on the three types of compensation components in the trajectory compensation map to obtain a precise marking area includes:

[0036] Based on the three types of compensation components in the trajectory compensation diagram, the compensation parameters are determined;

[0037] Based on the compensation parameters, the auxiliary laser parameters of the auxiliary laser system are corrected to obtain the corrected auxiliary laser parameters;

[0038] Based on the corrected auxiliary laser parameters, lines are drawn to obtain a corrected region image; error detection is performed on the corrected region image to obtain the error value;

[0039] If the error value is less than the error threshold, then the precise line-drawing area is obtained;

[0040] If the error value is equal to or greater than the error threshold, a secondary parameter correction is performed until the corrected error value is less than the error threshold, and then the precise line-drawing area is obtained.

[0041] Secondly, embodiments of this application provide a high-precision perovskite scribing system with master-slave laser coordination. This system includes: an acquisition module, a detection module, a generation module, and a correction module.

[0042] The acquisition module is used to perform preliminary scribing of the battery assembly to be cut according to a preset fixed trajectory through the main laser system, so as to obtain the preliminary scribing area.

[0043] The detection module is used to perform real-time tracking detection on the preliminary scribing area through an optical detection system to obtain detection data; and based on the detection data, to determine the depth difference map and width difference map of the preliminary scribing.

[0044] The generation module is used to generate a trajectory compensation map based on the depth difference map and the width difference map through a preset neural network; wherein, each pixel in the trajectory compensation map contains three types of compensation components: laser power, moving speed, and spot size;

[0045] The correction module is used to correct the auxiliary laser parameters of the auxiliary laser system based on the three types of compensation components in the trajectory compensation map, so as to obtain a precise line drawing area.

[0046] Thirdly, embodiments of this application provide a high-precision perovskite scribing device with master-slave laser synergy, comprising: a processor and a memory; wherein,

[0047] The memory is used to store computer programs;

[0048] The processor is configured to call and run the computer program from the memory to perform the method as described in the first aspect.

[0049] Fourthly, embodiments of this application provide a computer-readable storage medium storing executable instructions for causing a processor to perform the method described in the first aspect.

[0050] This application provides a method and system for high-precision scribing of perovskite using a master-auxiliary laser synergy. The method includes: using a master laser system to initially scribble a pre-defined fixed trajectory for a battery assembly to be cut, obtaining a preliminary scribing area; using an optical detection system to perform real-time tracking detection on the preliminary scribing area, obtaining detection data; and based on the detection data, determining a depth difference map and a width difference map of the preliminary scribing; generating a trajectory compensation map using a pre-defined neural network based on the depth difference map and the width difference map; wherein each pixel in the trajectory compensation map contains three types of compensation components: laser power, moving speed, and spot size; and correcting the auxiliary laser parameters of the auxiliary laser system based on the three types of compensation components in the trajectory compensation map to obtain a precise scribing area. In the above scheme, a closed-loop system of "real-time detection - intelligent compensation - precise correction" is constructed, achieving automated elimination of scribing errors through three steps. The optical system tracks and detects in real time, accurately capturing multi-dimensional errors such as scribing depth, width, and burrs. A preset neural network converts the error data into pixel-level compensation parameters (power, speed, and spot size). The auxiliary laser is synchronized with the main laser in time and space, and the error is corrected in a targeted manner according to the compensation parameters (i.e., three types of compensation components). This upgrades the traditional open-loop processing to closed-loop control, taking into account accuracy, efficiency, and compatibility, improving scribing accuracy and solving the problem of traditional accuracy loss of control. Attached Figure Description

[0051] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0052] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0053] Figure 1 A flowchart illustrating a high-precision perovskite scribing method using a master-slave laser synergy, provided in an embodiment of this application;

[0054] Figure 2 A schematic diagram of a high-precision scribing method for perovskites using a master-auxiliary laser synergy, provided as an embodiment of this application;

[0055] Figure 3 A schematic diagram of actual depth data for a high-precision perovskite scribing method using a master-slave laser synergy provided in an embodiment of this application;

[0056] Figure 4 A schematic diagram of a high-precision perovskite scribing system with master-auxiliary laser synergy provided in this application embodiment;

[0057] Figure 5 This is a schematic diagram of a high-precision perovskite scribing device with primary and secondary laser synergy, provided as an embodiment of this application. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0059] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this application is for the purpose of describing embodiments of this application only and is not intended to be limiting of this application.

[0060] In the following description, references to "some embodiments," "this embodiment," "this application embodiment," and examples, etc., describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subset of all possible embodiments and may be combined with each other without conflict.

[0061] If the application documents contain similar descriptions such as "first / second", the following explanation shall be added: In the following description, the terms "first / second / third" are used only to distinguish similar objects and do not represent a specific order of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0062] Current traditional laser scribing technology is difficult to adapt to the special characteristics of perovskite processing, and has the following problems:

[0063] First, the problem of errors caused by traditional offline inspection is difficult to remedy. Traditional processes require offline equipment to detect errors after the main laser has completed the scribing. If the substrate has problems such as insufficient depth or insufficient width, it cannot be corrected in real time or it needs to be scrapped directly, resulting in high material loss and making it difficult to meet the cost reduction requirements of industrialization.

[0064] Secondly, the problem of insufficient scribing accuracy caused by the extensive compensation strategy is that traditional compensation often adopts "uniform parameter adjustment" (such as changing the laser energy as a whole), without developing an adaptation scheme for different error types in different areas (such as depth deviation and insufficient width). Especially when facing local deformation of large-format substrates, it is easy to "correct one place and damage multiple places", which cannot meet the high-precision scribing requirements of perovskite modules.

[0065] Third, single-laser processing struggles to balance the contradiction between efficiency and correction effectiveness. Traditionally, it relies solely on the main laser to complete "processing + correction": if the processing speed is reduced to ensure accuracy, it will lead to low efficiency; if the speed is increased to pursue cycle time, it will exacerbate errors due to insufficient parameter adaptation, making it impossible to meet the core industrial demands of "high precision" and "high mass production efficiency".

[0066] Fourth, the poor equipment adaptability leads to multi-scenario compatibility problems. Traditional laser scribing equipment requires manual readjustment of core parameters for different substrate materials (such as glass and flexible substrates) or different film layers (such as conductive layers, perovskite layers, and top electrode layers). The changeover cycle is long and the film layer is easily damaged due to improper adjustment. It has poor compatibility and is difficult to adapt to the production needs of multiple types of perovskite components.

[0067] Based on this, embodiments of this application provide a high-precision perovskite scribing method using master-slave laser synergy. Figure 1 This is a flowchart illustrating a high-precision perovskite scribing method using master-slave laser synergy, provided in an embodiment of this application. Figure 1 The steps shown are explained.

[0068] S101. Using the main laser system, preliminary lines are drawn on the battery assembly to be cut according to a preset fixed trajectory to obtain the preliminary drawn area.

[0069] In some embodiments of this application, the main laser system performs preliminary scribing of the battery assembly to be cut according to a preset fixed trajectory. During the scribing process, the laser intensity and core processing parameters remain fixed. The core processing parameters include laser power, laser moving speed, spot size, laser frequency, and other intensity-related and motion-related parameters.

[0070] In some embodiments of this application, a high-precision scribing method for perovskite solar cells using a master-slave laser synergy is adapted to the perovskite solar cell scenario.

[0071] In some embodiments of this application, a high-precision perovskite scribing method with master-slave laser synergy is adapted to a high-precision perovskite scribing system with master-slave laser synergy.

[0072] S102. The preliminary marking area is monitored in real time using an optical inspection system to obtain inspection data; and based on the inspection data, the depth difference map and width difference map of the preliminary marking are determined.

[0073] In some embodiments of this application, the preliminary marking area is monitored in real time by an optical detection system to obtain detection data; the accuracy of the preliminary marking is determined based on the detection data, namely the depth difference map and width difference map of the preliminary marking.

[0074] In some embodiments of this application, the preliminary line-marking area is continuously photographed using an optical detection system to obtain the original image of the preliminary line-marking area; the original image is subjected to image binarization processing, three-dimensional visual reconstruction, and high-frequency component extraction to obtain detection data; wherein, the detection data includes actual line width data and actual depth data; based on the actual depth data and actual line width data, a pixel-by-pixel comparison is performed with a preset target depth and a preset target line width, and the difference is calculated based on the pixel comparison results to obtain a depth difference map and a width difference map.

[0075] S103. Based on the depth difference map and the width difference map, a trajectory compensation map is generated through a preset neural network; wherein, each pixel in the trajectory compensation map contains three types of compensation components: laser power, moving speed, and spot size.

[0076] In some embodiments of this application, the detection data further includes a spur distribution map. The preset neural network includes a trajectory sub-neural network, a light intensity sub-neural network, and a correction sub-neural network.

[0077] In some embodiments of this application, a trajectory sub-neural network is used to extract features from the width difference map and the spur distribution map to obtain a first moving speed feature map and a first spot size feature map; a light intensity sub-neural network is used to extract features from the depth difference map, the width difference map, and the spur distribution map to obtain a second moving speed feature map, a second spot size feature map, and a laser power feature map; a correction sub-neural network is used to perform feature fusion and parameter regression on the first moving speed feature map, the first spot size feature map, the second moving speed feature map, the second spot size feature map, and the laser power feature map to obtain the laser power, moving speed, and spot size; and a trajectory compensation map is determined based on the laser power, moving speed, and spot size.

[0078] For example, the depth difference map and width difference map are used as the core input of the neural network, combined with the spur distribution map (as an auxiliary feature). Through parallel feature extraction of the trajectory sub-neural network and the light intensity sub-neural network, and feature fusion and parameter regression of the correction sub-neural network, the final output is a trajectory compensation map with the same size as the scribing area. Each pixel of the trajectory compensation map corresponds to three types of compensation components, which are used to adjust the processing parameters of the auxiliary laser.

[0079] S104. Based on the three types of compensation components in the trajectory compensation diagram, the auxiliary laser parameters of the auxiliary laser system are corrected to obtain a precise scribing area.

[0080] In some embodiments of this application, compensation parameters are determined based on three types of compensation components in the trajectory compensation map; based on the compensation parameters, the auxiliary laser parameters of the auxiliary laser system are corrected to obtain corrected auxiliary laser parameters; based on the corrected auxiliary laser parameters, lines are drawn to obtain a corrected region image; error detection is performed on the corrected region image to obtain an error value; if the error value is less than the error threshold, a precise drawn region is obtained; if the error value is equal to or greater than the error threshold, a second parameter correction is performed until the corrected error value is less than the error threshold, and then a precise drawn region is obtained.

[0081] For example, using the pixel-level compensation components of the trajectory compensation map as the control basis, the initial error is accurately corrected through the auxiliary laser system. The core is to achieve error closed-loop elimination through "parameter adaptation - synchronous control - adaptive processing".

[0082] Understandably, to address the pain points of traditional laser scribing—poor accuracy, low efficiency, and weak adaptability—a closed-loop system of "real-time detection - intelligent compensation - precise correction" is constructed to automatically eliminate scribing errors through three steps. The optical system tracks and detects in real time, accurately capturing multi-dimensional errors such as scribing depth, width, and burrs; a dual-network collaborative neural network transforms error data into pixel-level compensation parameters (power, speed, and spot size); the auxiliary laser is synchronized with the main laser in time and space, and errors are corrected specifically according to the compensation parameters. This upgrades traditional open-loop processing to closed-loop control, balancing accuracy, efficiency, and compatibility, improving scribing accuracy, and solving the problem of uncontrolled accuracy in traditional methods.

[0083] In some embodiments of this application, S102 can be implemented by S201-S203, as follows:

[0084] S201. The preliminary line-marked area is continuously photographed using an optical detection system to obtain the original image of the preliminary line-marked area.

[0085] S202. Perform image binarization, 3D visual reconstruction, and high-frequency component extraction on the original image to obtain detection data; the detection data includes actual line width data and actual depth data.

[0086] In some embodiments of this application, the detection data includes actual line width data, actual depth data, and burr distribution map.

[0087] In some embodiments of this application, the original image is binarized to obtain a binarized image; edge extraction and feature analysis are performed on the binarized image to obtain actual line width data; the original image is reconstructed using three-dimensional vision processing technology to obtain a depth image; pixel-level depth values ​​are calculated on the depth image to determine the actual depth data; and high-frequency component extraction and feature recognition are performed on the original image using frequency domain filtering to obtain a burr distribution map.

[0088] In some embodiments of this application, high-frequency components are extracted from the original image by frequency domain filtering to obtain high-frequency components; threshold segmentation processing is performed on the high-frequency components by multiple preset thresholds to obtain a high-frequency truncated image; edge burr feature recognition is performed on the high-frequency truncated image to obtain a burr distribution map.

[0089] S203. Based on the actual depth data and actual line width data, perform pixel-by-pixel comparison with the preset target depth and preset target line width respectively, and calculate the difference based on the pixel comparison results to obtain the depth difference map and the width difference map.

[0090] For example, the preliminary marking area can be monitored in real time using an optical inspection system to obtain inspection data; and based on the inspection data, the depth difference map and width difference map of the preliminary marking can be determined through the following steps:

[0091] S21. The initial line-marking area is continuously photographed using an optical detection system to obtain the original image of the initial line-marking area; among which, considering image noise, the photographed image can be first subjected to Gaussian filtering, and the image after Gaussian filtering is the original image.

[0092] S22. Based on the original image, generate a binary image through image binarization; perform edge extraction and feature analysis on the binary image to extract the actual line width data.

[0093] For example, during binarization, a threshold can be used, with values ​​above the threshold being 1 and values ​​below the threshold being 0; during edge extraction, the edge image can be calculated using the Canny operator; and during feature analysis, the actual line width can be obtained by performing pixel-level statistics on the edge image.

[0094] For example, the actual line width is as follows Figure 2 As shown, d1 between K1 and K2 is the actual line width, d2 between K3 and K4 is the actual line width, and the line width between P1 and P2 is the target line width, with d1=d2, meaning the target line width is greater than the actual line width.

[0095] S23. Based on the original image, a depth image is reconstructed using three-dimensional vision processing technology; pixel-level depth values ​​are calculated on the depth image to determine the actual depth data; wherein, the three-dimensional vision processing technology can be visible light-based binocular vision, or TOF, or structured light.

[0096] For example, actual depth data such as Figure 3 As shown, the distance between T1 and T2 is the target depth, and the distance between N1 and N2 is the actual depth. The target depth is greater than the actual depth.

[0097] S24. Based on the original image, high-frequency components are extracted through frequency domain filtering. Threshold segmentation is then performed on the extracted high-frequency components to generate a high-frequency truncated image. Edge burr feature recognition is then performed on the high-frequency truncated image to extract the burr distribution map. The specific steps are as follows:

[0098] High-frequency components are extracted using the Laplacian operator;

[0099] Based on multiple pre-set thresholds, high-frequency components are CLIPed to extract the high-frequency components corresponding to the glitch.

[0100] The high-frequency components that are not punctured are set to 0, while the high-frequency components that are punctured are kept at their original values, thus obtaining the punctured distribution map.

[0101] S25. The extracted actual depth data and actual line width data are compared pixel by pixel with the preset target depth and target line width, respectively. The depth difference map and width difference map are generated by calculating the difference. Specifically, as shown in formulas (1) and (2).

[0102] (1)

[0103] (2)

[0104] in, This is a depth difference map. This is a width difference graph. D ACT W represents the pixels of the actual depth data. ACT The actual width data in pixels. D TAR For the pixels of the target depth data, W TAR The target width data in pixels.

[0105] In some embodiments of this application, S103 can be implemented by S301-S304, as follows:

[0106] S301. Using a trajectory sub-neural network, feature extraction is performed on the width difference map and the burr distribution map to obtain the first moving speed feature map and the first spot size feature map.

[0107] S302. Using a light intensity sub-neural network, feature extraction is performed on the depth difference map, width difference map, and burr distribution map to obtain the second moving speed feature map, the second spot size feature map, and the laser power feature map.

[0108] S303. By using the corrected sub-neural network, feature fusion and parameter regression are performed on the first moving speed feature map, the first spot size feature map, the second moving speed feature map, the second spot size feature map, and the laser power feature map to obtain the laser power, moving speed, and spot size.

[0109] In some embodiments of this application, a corrected sub-neural network is used to fuse the first moving speed feature map and the second moving speed feature map to obtain a final moving speed feature map; a corrected sub-neural network is used to fuse the first spot size feature map and the second spot size feature map to obtain a final spot size feature map; the final moving speed feature map, the final spot size feature map, and the laser power feature map are spliced ​​to obtain pixel fusion features; and a fully connected layer and a linear function are used to perform parameter regression on the pixel fusion features to obtain the laser power, moving speed, and spot size.

[0110] S304. Determine the trajectory compensation diagram based on the laser power, moving speed, and spot size.

[0111] For example, the depth difference map and width difference map are used as the core inputs of the neural network, combined with the spur distribution map (as an auxiliary feature). Through parallel feature extraction by the trajectory sub-neural network and the light intensity sub-neural network, and feature fusion and parameter regression by the correction sub-neural network, the final output is a trajectory compensation map with the same size as the drawn area. Each pixel of the trajectory compensation map corresponds to three types of compensation components. The specific steps are as follows:

[0112] S31. The input to the trajectory sub-neural network is the spur distribution map and the width difference map, and the output is the first moving speed feature map and the first spot size feature map. The specific network calculation steps are as follows:

[0113] a. Perform channel stitching on the burr distribution map and width difference map, and then perform min-max normalization. The normalization formula is: , where F is the pixel value in any image, and max(F) and min(F) are the maximum and minimum pixel values ​​in the corresponding image, respectively;

[0114] b. Use 2 layers of 3×3 ordinary convolution (stride 1) + ReLU activation to extract spatial features;

[0115] c. Spatial features are reduced in dimensionality by 4×4 convolution, outputting the first moving velocity feature map V1 and the first spot size feature map S1.

[0116] S32. The input to the optical intensifier neural network is a glitch map, a depth difference map, and a width difference map. The output is a second moving velocity feature map, a second spot size feature map, and a laser power feature map. The specific steps are as follows:

[0117] d. Perform channel stitching on the burr distribution map, depth difference map, and width difference map, and then perform min-max normalization.

[0118] e. Use 4 layers of 5×5 ordinary convolution (stride 4) + ReLU activation to extract light intensity correlation features;

[0119] f. The light intensity correlation features are reduced in dimensionality by 2×2 convolution, and the output is the second moving speed feature map V2, the second spot size feature map S2, and the laser power feature map P.

[0120] S33. The input to the correction sub-neural network is the first moving speed feature map, the second moving speed feature map, the first spot size feature map, the second spot size feature map, and the laser power feature map. The output is the corresponding laser power, moving speed, and spot size. The specific steps are as follows:

[0121] g. Fusion of homologous features: V1 and V2 are fused, and S1 and S2 are fused. The fusion formula is as follows:

[0122] (3)

[0123] (4)

[0124] h. Feature concatenation: The V, S, and P components are concatenated at corresponding positions to obtain the pixel fusion feature map at that position.

[0125] i. Parametric Regression: The fused features are mapped to four-channel compensation values ​​through a fully connected layer and a linear function. The calculation method is shown in formula (5), where, k X is the slope, O is the input, O is the compensation, and Y is the output. Y V , Y S , Y P The three compensation components in the trajectory compensation diagram are laser power, moving speed, and spot size.

[0126] (5)

[0127] It should be noted that the three types of compensation components are the compensation parameters.

[0128] After obtaining the three types of compensation components in the trajectory compensation map, perform the following steps:

[0129] S41, the calculation method for each pixel of the auxiliary laser actual power, auxiliary laser actual moving speed, and auxiliary laser actual spot size is shown in formula (6), where base is the auxiliary laser parameter. Y V1 , Y S1 , Y P1 These are the corrected auxiliary laser parameters.

[0130] (6)

[0131] It should be noted that the corrected auxiliary laser parameters are the actual auxiliary laser power, the actual auxiliary laser moving speed, and the actual auxiliary laser spot size.

[0132] S42. During the correction process, the image of the correction area is acquired in real time by a coaxial optical sensor, and the correction effect is verified by the difference comparison operator. If the error is greater than the threshold, secondary compensation is triggered until the error meets the preset accuracy requirements.

[0133] The formula for the difference comparison operator is:

[0134] (7)

[0135] in, D thr , W thr The errors in the depth difference map and the width difference map exceeded the threshold, respectively.

[0136] Understandably, firstly, it significantly improves scribing accuracy, solving the problem of uncontrolled accuracy in traditional methods. This application utilizes a full-link technology of "real-time optical pixel-by-pixel detection + neural network pixel-level compensation + auxiliary laser precision correction" to achieve precise quantification and targeted elimination of scribing errors. Whether it's depth deviation, uneven line width, trajectory offset, or edge burrs, all can be effectively controlled through layered detection and adaptive correction, completely solving the problem of unstable accuracy caused by substrate deformation and light intensity fluctuations in traditional laser scribing, meeting the stringent requirements for scribing accuracy in precision devices such as perovskite solar cells. Secondly, it optimizes the processing flow, reducing costs and material waste. Compared to the traditional "offline detection-rework correction" mode, this application achieves integrated operation of "scibing-detection-correction": the optical system and main laser follow the detection in real time, and the auxiliary laser and main laser correct synchronously in time and space, eliminating the need for downtime or offline processing, significantly shortening the overall processing cycle; at the same time, through a secondary compensation verification mechanism, it effectively reduces the scribing defect rate, reduces substrate material waste and rework costs, and adapts to the efficiency and cost requirements of industrial mass production. Thirdly, it has strong adaptability and compatibility, covering the needs of multiple scenarios. This application relies on a hardware adaptation mapping and error level partitioning correction mechanism to flexibly adapt to the scribing requirements of different substrate materials (such as glass and flexible substrates) and different film layer structures (such as conductive layers, perovskite composite layers, and top electrode layers). It can automatically adjust auxiliary laser parameters and correction strategies for different types of errors (insufficient depth, trajectory deviation, etc.) without requiring equipment replacement or re-tuning, and is compatible with scribing error correction for all processes from P1 to P3, reducing equipment investment and replacement costs. Fourth, the technology has strong practical applicability, promoting industrialization. The neural network structure used in this invention is simplified and computationally efficient, adapting to industrial-grade real-time processing requirements; the optical inspection and auxiliary laser system can be upgraded based on existing laser scribing equipment without requiring new development, lowering the technology implementation threshold. The overall solution achieves automated operation through closed-loop control, reducing manual intervention and adapting to large-format, high-cycle industrial production scenarios, providing support for the large-scale application of precision laser processing technology.

[0137] Based on the above embodiments, this application also provides a high-precision perovskite scribing system using master-slave laser synergy. Figure 4 As shown, Figure 4This is a schematic diagram of a high-precision perovskite scribing system with master-slave laser synergy provided in an embodiment of this application. The system 4 includes: an acquisition module 401, a detection module 402, a generation module 403, and a correction module 404.

[0138] The acquisition module 401 is used to perform preliminary scribing of the battery assembly to be cut according to a preset fixed trajectory through the main laser system to obtain the preliminary scribing area.

[0139] The detection module 402 is used to perform real-time tracking detection on the preliminary scribing area through an optical detection system to obtain detection data; and based on the detection data, to determine the depth difference map and width difference map of the preliminary scribing.

[0140] The generation module 403 is used to generate a trajectory compensation map based on the depth difference map and the width difference map through a preset neural network; wherein, each pixel in the trajectory compensation map contains three types of compensation components: laser power, moving speed, and spot size;

[0141] The correction module 404 is used to correct the auxiliary laser parameters of the auxiliary laser system based on the three types of compensation components in the trajectory compensation map, so as to obtain a precise line drawing area.

[0142] Based on the above embodiments, this application also provides a high-precision perovskite scribing device using a master-slave laser synergy. Figure 5 As shown, Figure 5 This is a schematic diagram of a high-precision perovskite scribing device with master-slave laser synergy provided in an embodiment of this application. The device includes a processor 501 and a memory 502. The memory 502 stores a computer program; the processor 501 retrieves and runs the computer program from the memory to execute the high-precision perovskite scribing method with master-slave laser synergy as described in the above embodiment.

[0143] In the embodiments of this application, the processor 501 described above can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that for different devices, the electronic device used to implement the above processor function can also be other types, and the embodiments of this application do not specifically limit it.

[0144] This application provides a computer-readable storage medium storing a computer program for implementing, when executed by a processor, a high-precision perovskite scribing method with master-slave laser coordination as described in any of the above embodiments.

[0145] For example, the program instructions corresponding to the master-slave laser coordinated high-precision perovskite scribing method in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the program instructions corresponding to the master-slave laser coordinated high-precision perovskite scribing method in the storage media are read or executed by an electronic device, the master-slave laser coordinated high-precision perovskite scribing method as described in any of the above embodiments can be realized.

[0146] Furthermore, in the embodiments of this application, the functional modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.

[0147] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, 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. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0148] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the embodiments in this application are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, these will not be repeated here.

[0149] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0150] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0151] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0152] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0153] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0154] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0155] The above description is merely an embodiment of this application, but the protection scope of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A high-precision perovskite scribing method using master-slave laser synergy, characterized in that, The method includes: The main laser system performs preliminary scribing of the battery module to be cut according to a preset fixed trajectory, thus obtaining the preliminary scribing area. The preliminary marking area is monitored in real time using an optical inspection system to obtain inspection data; and based on the inspection data, the depth difference map and width difference map of the preliminary marking are determined. Based on the depth difference map and the width difference map, a trajectory compensation map is generated through a preset neural network; wherein, each pixel in the trajectory compensation map contains three types of compensation components: laser power, moving speed, and spot size; Based on the three types of compensation components in the trajectory compensation diagram, the auxiliary laser parameters of the auxiliary laser system are corrected to obtain a precise line drawing area. The preset neural network includes a trajectory sub-neural network, a light intensity sub-neural network, and a correction sub-neural network; the generation of a trajectory compensation map based on the depth difference map and the width difference map through the preset neural network includes: The trajectory sub-neural network is used to extract features from the width difference map and the spur distribution map to obtain a first moving speed feature map and a first spot size feature map. The optical intensity sub-neural network is used to extract features from the depth difference map, the width difference map, and the burr distribution map to obtain a second moving speed feature map, a second spot size feature map, and a laser power feature map. The corrective sub-neural network is used to perform feature fusion and parameter regression on the first moving speed feature map, the first spot size feature map, the second moving speed feature map, the second spot size feature map, and the laser power feature map to obtain the laser power, the moving speed, and the spot size. The trajectory compensation map is determined based on the laser power, the moving speed, and the spot size.

2. The method according to claim 1, characterized in that, The process involves real-time tracking and detection of the initially marked area using an optical detection system to obtain detection data; and based on the detection data, determining the depth difference map and width difference map of the initially marked area, including: The optical detection system continuously captures images of the initially marked area to obtain the original image of the initially marked area. The original image is subjected to image binarization, 3D visual reconstruction, and high-frequency component extraction to obtain the detection data; wherein, the detection data includes actual line width data and actual depth data; Based on the actual depth data and the actual line width data, a pixel-by-pixel comparison is performed with the preset target depth and the preset target line width, and the difference is calculated based on the pixel comparison results to obtain the depth difference map and the width difference map.

3. The method according to claim 2, characterized in that, The detection data also includes a burr distribution map; The detection data is obtained by performing image binarization, 3D visual reconstruction, and high-frequency component extraction on the original image, including: The original image is binarized to obtain a binarized image; and edge extraction and feature analysis are performed on the binarized image to obtain the actual line width data. The original image is reconstructed using 3D vision processing technology to obtain a depth image; and pixel-level depth values ​​are calculated from the depth image to determine the actual depth data. By using frequency domain filtering, high-frequency components are extracted and features are identified from the original image to obtain the spiky distribution map.

4. The method according to claim 3, characterized in that, The step of extracting high-frequency components and identifying features from the original image through frequency domain filtering to obtain the spur distribution map includes: The high-frequency components are extracted from the original image by frequency domain filtering. By using multiple pre-set thresholds, the high-frequency components are subjected to threshold segmentation to obtain a high-frequency truncated image; Edge burr feature recognition is performed on the high-frequency truncated image to obtain the burr distribution map.

5. The method according to claim 1, characterized in that, The step of performing feature fusion and parameter regression on the first moving speed feature map, the first spot size feature map, the second moving speed feature map, the second spot size feature map, and the laser power feature map through the corrected sub-neural network to obtain the laser power, the moving speed, and the spot size includes: The first and second movement speed feature maps are fused using the corrective sub-neural network to obtain the final movement speed feature map. The first spot size feature map and the second spot size feature map are fused using the corrective sub-neural network to obtain the final spot size feature map. The final moving speed feature map, the final spot size feature map, and the laser power feature map are spliced ​​together to obtain pixel fusion features; By using a fully connected layer and a linear function, parameter regression is performed on the pixel fusion features to obtain the laser power, the moving speed, and the spot size.

6. The method according to claim 1, characterized in that, The auxiliary laser parameters of the auxiliary laser system are corrected based on the three types of compensation components in the trajectory compensation map to obtain a precise line drawing area, including: Based on the three types of compensation components in the trajectory compensation diagram, the compensation parameters are determined; Based on the compensation parameters, the auxiliary laser parameters of the auxiliary laser system are corrected to obtain the corrected auxiliary laser parameters; Based on the corrected auxiliary laser parameters, lines are drawn to obtain a corrected region image; error detection is performed on the corrected region image to obtain the error value; If the error value is less than the error threshold, then the precise line-drawing area is obtained; If the error value is equal to or greater than the error threshold, a secondary parameter correction is performed until the corrected error value is less than the error threshold, and then the precise line-drawing area is obtained.

7. A high-precision perovskite scribing system with master-slave laser synergy, characterized in that, The master-slave laser coordinated perovskite high-precision scribing system includes: an acquisition module, a detection module, a generation module, and a correction module, wherein, The acquisition module is used to perform preliminary scribing of the battery assembly to be cut according to a preset fixed trajectory through the main laser system, so as to obtain the preliminary scribing area. The detection module is used to perform real-time tracking detection on the preliminary scribing area through an optical detection system to obtain detection data; and based on the detection data, to determine the depth difference map and width difference map of the preliminary scribing. The generation module is used to generate a trajectory compensation map based on the depth difference map and the width difference map using a preset neural network. Each pixel in the trajectory compensation map contains three compensation components: laser power, moving speed, and spot size. The preset neural network includes a trajectory sub-neural network, a light intensity sub-neural network, and a correction sub-neural network. Generating the trajectory compensation map based on the depth difference map and the width difference map using the preset neural network includes: extracting features from the width difference map and the burr distribution map using the trajectory sub-neural network to obtain a first moving speed feature map and a first spot size feature map; extracting features from the depth difference map, the width difference map, and the burr distribution map using the light intensity sub-neural network to obtain a second moving speed feature map, a second spot size feature map, and a laser power feature map; performing feature fusion and parameter regression on the first moving speed feature map, the first spot size feature map, the second moving speed feature map, the second spot size feature map, and the laser power feature map using the correction sub-neural network to obtain the laser power, the moving speed, and the spot size; and determining the trajectory compensation map based on the laser power, the moving speed, and the spot size. The correction module is used to correct the auxiliary laser parameters of the auxiliary laser system based on the three types of compensation components in the trajectory compensation map, so as to obtain a precise line drawing area.

8. A high-precision perovskite scribing device with master-auxiliary laser synergy, characterized in that, include: Processor and memory, of which, The memory is used to store computer programs; The processor is configured to call and run the computer program from the memory to perform the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, It stores executable instructions for causing a processor to execute, thereby implementing the method of any one of claims 1 to 6.

Images