A high-precision segmentation method and system for a PCB board splitter based on intelligent control

Abstract

The present disclosure provides a PCB splitter high-precision segmentation method and system based on intelligent control, relating to the technical field of data processing, which comprises: obtaining a preset position and laying out an image acquisition device to obtain the laying-out result of the image acquisition device; obtaining a target PCB physical object; writing a cycle moving scanning array through the motion module to obtain a target cycle scanning array; performing cycle image acquisition on the target PCB physical object to obtain image acquisition data; performing image processing on the image acquisition data to obtain target PCB image data; constructing a splitting numerical control model to obtain cutting trajectory data; and performing data conversion on the cutting trajectory data based on an XYZ Cartesian mechanical coordinate system to obtain mechanical cutting trajectory data. The technical problem of insufficient cutting efficiency and cutting precision of the splitter due to low intelligence and low precision of the splitter in the prior art is solved.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, specifically to a high-precision PCB depaneling method and system based on intelligent control. Background Technology

[0002] PCB (Printed Circuit Board) is an important electronic component. It serves as the support for electronic components and the carrier for their electrical interconnection. Because it is manufactured using electronic printing technology, it is called a "printed" circuit board.

[0003] A PCB depaneling machine is a machine used to connect PCB boards. Traditional manual folding methods suffer from uncontrollable quality and pose safety hazards for operators. Furthermore, uneven folding force and varying angles can lead to damage to PCB circuits, components, and solder baths; component detachment; copper foil deformation; and board detachment, resulting in shortened lifespan and other defects. The emergence of PCB depaneling machines overcomes these quality issues compared to traditional manual folding.

[0004] Currently, most domestic PCB splitting systems use CCD-assisted human eye teaching or DXF file import for programming. CCD teaching is prone to causing visual fatigue, leading to misreading cutting positions, and requires significant programming time, hindering rapid line changeovers on production lines. DXF file import requires customers to provide a suitable version of the GERBER file and use CAD or similar drawing software to draw the trajectory before generating the cutting file. This method allows for off-line programming, but requires third-party drawing software, demanding certain drawing skills from operators, limiting its applicability. Furthermore, after installation, calibration points need to be imported to match the machine's coordinate system.

[0005] Currently, existing technologies suffer from insufficient cutting efficiency and precision due to the low level of intelligence and precision of PCB depaneling machines. Summary of the Invention

[0006] This disclosure provides a high-precision PCB depaneling method and system based on intelligent control, which solves the technical problems in the prior art where the depaneling machine's low level of intelligence and low precision leads to insufficient cutting efficiency and precision.

[0007] According to a first aspect of this disclosure, a high-precision PCB depaneling method based on intelligent control is provided, comprising: establishing an XYZ Cartesian mechanical coordinate system; obtaining a preset position based on the XYZ Cartesian mechanical coordinate system, and deploying an image acquisition device based on the preset position to obtain an image acquisition device deployment result; obtaining a target PCB physical object; based on the target PCB physical object, programming a cyclic scanning array through a motion module to obtain a target cyclic scanning array; based on the image acquisition device deployment result and the target cyclic scanning array, controlling the image acquisition device through the motion module to perform cyclic image acquisition on the target PCB physical object to obtain image acquisition data; performing image processing on the image acquisition data to obtain target PCB image data; constructing a depaneling CNC model, inputting the target PCB image data into the depaneling CNC model to obtain cutting trajectory data; and performing data conversion on the cutting trajectory data based on the XYZ Cartesian mechanical coordinate system to obtain mechanical cutting trajectory data.

[0008] According to a second aspect of this disclosure, a high-precision PCB depaneling system based on intelligent control is provided, comprising: a coordinate system construction module for constructing an XYZ Cartesian mechanical coordinate system; an image acquisition device deployment module for obtaining a preset position based on the XYZ Cartesian mechanical coordinate system and deploying an image acquisition device based on the preset position to obtain an image acquisition device deployment result; a PCB physical object acquisition module for obtaining a target PCB physical object; and a target cyclic scanning array acquisition module for obtaining a target cyclic scanning array based on the target PCB physical object by programming a cyclic moving scanning array through a motion module. The system includes: a cyclic image acquisition module, which, based on the layout of the image acquisition device and the target cyclic scanning array, controls the image acquisition device via the motion module to perform cyclic image acquisition on the target PCB to obtain image acquisition data; an image processing module, which processes the image acquisition data to obtain target PCB image data; a PCB CNC model construction module, which constructs a PCB CNC model and inputs the target PCB image data into the PCB CNC model to obtain cutting trajectory data; and a data conversion module, which converts the cutting trajectory data based on the XYZ Cartesian mechanical coordinate system to obtain mechanical cutting trajectory data.

[0009] According to a third aspect of this disclosure, an electronic device is provided, comprising:

[0010] At least one processor; and

[0011] A memory communicatively connected to the at least one processor; wherein,

[0012] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0013] According to the high-precision segmentation method for PCB depaneling machines based on intelligent control adopted in this disclosure, an XYZ Cartesian mechanical coordinate system is established; based on the XYZ Cartesian mechanical coordinate system, a preset position is obtained, and an image acquisition device is deployed based on the preset position to obtain the image acquisition device deployment result; a target PCB physical object is obtained; based on the target PCB physical object, a cyclic moving scan array is programmed through the motion module to obtain a target cyclic scan array; based on the image acquisition device deployment result and the target cyclic scan array, the motion module controls the image acquisition device to perform cyclic image acquisition on the target PCB physical object to obtain image acquisition data; the image acquisition data is processed to obtain target PCB image data. According to the method described above, a PCB splitting CNC model is constructed, and the target PCB image data is input into the PCB splitting CNC model to obtain cutting trajectory data. Based on the XYZ Cartesian mechanical coordinate system, the cutting trajectory data is converted to obtain mechanical cutting trajectory data. This disclosure obtains image acquisition data by performing cyclic array scanning on the target PCB, performing image processing on the image acquisition data, and obtaining a sample image training set through dynamic image stitching and fusion. The sample image training set is used as training samples to train the PCB splitting CNC model based on a convolutional neural network to obtain a PCB splitting CNC model that has been trained to convergence, and outputs the features of the position to be cut and feature points, thereby determining the PCB cutting trajectory data, achieving the technical effect of improving cutting efficiency and accuracy.

[0014] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0016] Figure 1 A flowchart illustrating a high-precision PCB depaneling method based on intelligent control, provided in an embodiment of this disclosure;

[0017] Figure 2 A schematic diagram of a high-precision PCB depaneling system based on intelligent control, provided in an embodiment of this disclosure;

[0018] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure.

[0019] Explanation of reference numerals in the attached diagram: Coordinate system setup module 11, Image acquisition device deployment module 12, PCB physical object acquisition module 13, Target cyclic scanning array acquisition module 14, Cyclic image acquisition module 15, Image processing module 16, Board splitting CNC model construction module 17, Data conversion module 18, Electronic device 800, Processor 801, Memory 802, Bus 803. Detailed Implementation

[0020] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0021] In order to solve the technical problems of insufficient cutting efficiency and precision of PCB depaneling machines due to low intelligence and low precision in existing technologies, the inventors of this disclosure have creatively developed a high-precision PCB depaneling method and system based on intelligent control.

[0022] Example 1

[0023] Figure 1 This application provides a high-precision PCB depaneling method based on intelligent control, which is applied to a high-precision PCB depaneling system based on intelligent control. The system includes an image acquisition device and a motion module, such as... Figure 1 As shown, the method includes:

[0024] Step S100: Establish the XYZ Cartesian machine coordinate system;

[0025] Specifically, the aforementioned PCB depaneling machine is used in electronic manufacturing for cutting printed circuit boards. It is suitable for PCB boards of various thicknesses. The aforementioned image acquisition device includes, but is not limited to, a machine vision system (CCD). An XYZ Cartesian mechanical coordinate system is established. The XYZ Cartesian mechanical coordinate system is the reference coordinate system for CNC machining. The Z-axis is the vertical position of the depaneling machine head, the X-axis is the horizontal direction, and the Y-axis is the horizontal direction. When the depaneling machine is cutting, the XYZ Cartesian mechanical coordinate system is used as the reference for cutting.

[0026] Step S200: Based on the XYZ Cartesian machine coordinate system, obtain a preset position, and based on the preset position, deploy the image acquisition device to obtain the image acquisition device deployment result;

[0027] Specifically, a preset position is obtained based on the XYZ Cartesian mechanical coordinate system. The preset position is the position of the head of the PCB depaneling machine. Based on the preset position, the image acquisition device is set up, and the image acquisition device setup result is obtained.

[0028] Step S300: Obtain the target PCB physical object;

[0029] Specifically, the target PCB is the PCB board that needs to be cut.

[0030] Step S400: Based on the target PCB physical object, the motion module is used to program the cyclic moving scan array to obtain the target cyclic scan array;

[0031] In this embodiment, based on the target PCB physical object, the motion module is used to program a cyclic scanning array to obtain the target cyclic scanning array. Step S400 further includes:

[0032] Step S410: The motion module includes multiple preset PCB parameters and multiple preset cyclic moving scan arrays, and the multiple preset PCB parameters and the multiple preset cyclic moving scan arrays have a corresponding relationship;

[0033] Step S420: Collect parameters from the target PCB physical object to obtain the target PCB parameters;

[0034] Step S430: Perform a correlation evaluation based on the target PCB parameters and the multiple preset PCB parameters to obtain the correlation evaluation results of multiple PCB parameters;

[0035] Step S440: Based on the correlation evaluation results of the multiple PCB parameters, the multiple preset cyclic moving scan arrays are screened to obtain the target cyclic scan array.

[0036] In this embodiment, step S440 further includes: Based on the correlation evaluation results of the multiple PCB parameters, the multiple preset cyclic moving scan arrays are screened to obtain the target cyclic scan array.

[0037] Step S441: Based on the correlation evaluation results of the multiple PCB parameters, determine the optimal correlation evaluation result;

[0038] Step S442: Based on the optimal correlation evaluation result, match the multiple preset cyclic moving scan arrays to obtain the target cyclic scan array.

[0039] Specifically, based on the target PCB, a cyclic scanning array is programmed through a motion module to obtain the target cyclic scanning array. The cyclic scanning array scans in a certain direction. If it reaches the edge position, it returns to the innermost layer to start the next round of scanning, and this action is repeated continuously.

[0040] Specifically, the aforementioned motion module includes multiple preset PCB parameters and multiple preset cyclic moving scan arrays. The preset PCB parameters are user-defined parameters such as the thickness and dimensions of the PCB board. The cyclic moving scan arrays scan in a specific direction; if they reach an edge, they return to the innermost layer to begin the next scan. These preset cyclic moving scan arrays are set based on the preset PCB parameters; therefore, there is a correspondence between the preset PCB parameters and the preset cyclic moving scan arrays—one preset PCB parameter corresponds to one preset cyclic moving scan array. Further, parameters are acquired from the target PCB object to obtain the target PCB parameters, which are the thickness, dimensions, and other parameters of the PCB object to be cut. A correlation evaluation is performed based on the target PCB parameters and the multiple preset PCB parameters to obtain the correlation evaluation results. In simpler terms, the target PCB parameters and the multiple preset PCB parameters are compared to determine their similarity. The correlation evaluation results represent the similarity between the target PCB parameters and the multiple preset PCB parameters. Based on these correlation evaluation results, the multiple preset cyclic moving scan arrays are then selected to obtain the target cyclic scanning array.

[0041] Specifically, based on the correlation evaluation results of multiple PCB parameters, the optimal correlation evaluation result is determined. The optimal correlation evaluation result is the one with the highest similarity among the multiple PCB parameter correlation evaluation results. Then, based on the optimal correlation evaluation result, multiple preset cyclic moving scan arrays are matched to determine the preset PCB parameters corresponding to the optimal correlation evaluation result. Each preset PCB parameter has its corresponding preset cyclic moving scan array. Based on the optimal correlation evaluation result, the corresponding preset cyclic moving scan array is matched among the multiple preset cyclic moving scan arrays to obtain the target cyclic scan array, which is the matched preset cyclic moving scan array.

[0042] Step S500: Based on the deployment results of the image acquisition device and the target cyclic scanning array, the motion module controls the image acquisition device to perform cyclic image acquisition on the target PCB to obtain image acquisition data;

[0043] In this embodiment, after obtaining the image acquisition data, step S500 further includes:

[0044] Step S510: Based on the image acquisition data, perform image quality assessment to obtain image quality assessment coefficients;

[0045] Step S520: Obtain image quality constraints;

[0046] Step S530: Determine whether the image quality evaluation coefficients meet the image quality constraints;

[0047] If the image quality evaluation coefficient does not meet the image quality constraint, an optimized image acquisition instruction is generated.

[0048] Step S540: Based on the optimized image acquisition command, obtain optimized image acquisition data of the target PCB, and update the image acquisition data based on the optimized image acquisition data.

[0049] Specifically, based on the deployment of the image acquisition device and the target cyclic scanning array, the motion module controls the image acquisition device to perform cyclic image acquisition on the target PCB to obtain image acquisition data.

[0050] Specifically, after acquiring image acquisition data, image quality assessment is performed based on this data to obtain an image quality assessment coefficient. Image quality assessment can evaluate various image features such as resolution, color depth, and image distortion. A simple scoring standard can be set to quantify the image quality assessment result; for example, a scale of 0-10 can be used, with scores closer to 10 indicating higher image quality. This evaluation of image quality yields the image quality assessment coefficient, which is the assessment result derived based on a certain evaluation standard. Further, image quality constraints are obtained; these constraints are a threshold set for the image quality assessment coefficient. The value range indicates that the image quality is acceptable within this image quality constraint range. You can set the value range according to the actual situation. Then, it is determined whether the image quality evaluation coefficient meets the image quality constraint. If the image quality evaluation coefficient does not meet the image quality constraint, it means that the image acquisition data is unacceptable. An optimized image acquisition command is generated. The optimized image acquisition command is the instruction issued to optimize the image acquisition. Based on the optimized image acquisition command, the optimized image acquisition data of the target PCB is obtained, and the image acquisition data is updated based on the optimized image acquisition data. It can be simply understood as replacing the image acquisition data with the optimized image acquisition data to complete the image acquisition data update.

[0051] Step S600: Perform image processing on the acquired image data to obtain target PCB image data;

[0052] In this embodiment, step S600 further includes performing image processing on the acquired image data to obtain target PCB image data:

[0053] Step S610: Denoise the image acquisition data based on the median filtering algorithm to obtain denoised image acquisition data;

[0054] Step S620: Perform image fusion based on the denoised image acquisition data to obtain the image fusion processing result;

[0055] Step S630: Perform image enhancement based on the image fusion processing result to obtain the target PCB image data.

[0056] After obtaining the image fusion processing result, step S620 in this embodiment of the application further includes:

[0057] Step S621: Based on the image fusion processing result, obtain feature location image data;

[0058] Step S622: Compare the feature location image data with the denoised image acquisition data to obtain the image comparison result;

[0059] Step S623: Based on the image comparison results, obtain the stitching and fusion accuracy information;

[0060] Step S624: Based on the stitching and fusion accuracy information, adjust the image fusion processing result by stitching and fusion to obtain the adjusted image fusion result.

[0061] Specifically, image processing is performed on the image dataset, including image denoising, image stitching and fusion, and image enhancement, to obtain target PCB image data, which serves as the reference image for subsequent cutting trajectory analysis.

[0062] Specifically, the image acquisition data is denoised using a median filtering algorithm to obtain denoised image acquisition data. Median filtering is a non-linear image smoothing method that effectively filters out impulse noise while preserving the edges of the target image. It can be implemented using MATLAB software. Further, the denoised image acquisition data is stitched and fused together. Image stitching, simply put, is combining several images with overlapping portions into a seamless panoramic or high-resolution image. These overlapping images may have been acquired at different times, from different viewpoints, or using different sensors. The output of image stitching is the union of the two input images. Image fusion is a process of removing gaps in stitched images. After image stitching, obvious seam marks and edge effects are often observed at the stitching boundaries. Due to lighting and color variations, the transitions at these boundaries are often poor. Image fusion is necessary to address this unnaturalness. Image stitching and fusion can be implemented using MATLAB software, allowing for the creation of image fusion processing results. Based on these results, image enhancement is performed to amplify useful information within the image. The goal of image enhancement is to improve the visual effect of the image, purposefully emphasizing overall or local characteristics, making previously unclear images clearer, amplifying the differences between features of different objects, suppressing uninteresting features, improving image quality, enriching information, and enhancing image interpretation and recognition capabilities, ultimately obtaining the target PCB image data.

[0063] Specifically, after obtaining the image fusion processing result, feature location image data is obtained based on the image fusion processing result. Feature location image data is the image data at the stitching and fusion point. The feature location image data is compared with the denoised image acquisition data to obtain image comparison results. Simply put, it compares whether there are differences in the position, size, shape, etc. of the stitching and fusion point, whether the stitching of the two images matches, and whether the stitching position is accurate. Based on the image comparison results, stitching and fusion accuracy information is obtained. The stitching and fusion accuracy information represents the stitching and fusion effect of the two images. The better the stitching and fusion effect, the higher the stitching and fusion accuracy. The stitching and fusion adjustment is performed on the image fusion processing result according to the stitching and fusion accuracy information to obtain the adjusted image fusion result. In other words, if the stitching and fusion accuracy is low, it means that the image stitching and fusion effect is not good. There may be problems such as deviation in the position of the stitching and fusion point or mismatch in the stitching position of the two images. It is necessary to adjust the stitching and fusion to make the position of the stitching and fusion point accurate and natural, thereby obtaining the adjusted image fusion result.

[0064] Step S700: Construct a PCB splitting CNC model, input the target PCB image data into the PCB splitting CNC model, and obtain cutting trajectory data;

[0065] In this embodiment, step S700 of constructing the CNC model for the split-plate assembly further includes:

[0066] Step S710: Obtain the model building database, wherein the model building database includes multiple sample PCB image data and multiple sample cutting trajectory data;

[0067] Step S720: Randomly divide the model construction database to obtain training dataset and test dataset;

[0068] Step S730: Connect to the modeling platform and construct the CNC model of the split board;

[0069] Step S740: Train and test the board CNC model based on the training dataset and the test dataset to obtain the board CNC model that meets the preset accuracy.

[0070] Specifically, the PCB splitting CNC model is trained based on a convolutional neural network. By inputting the target PCB image data, it outputs the features of the position to be cut and feature points, thereby determining the PCB cutting trajectory data.

[0071] Specifically, a model building database is obtained, which includes multiple sample PCB image data and multiple sample cutting trajectory data. The model building database is randomly divided to obtain training datasets and test datasets. For example, 60% of the data in the model building database can be used as the training dataset, and 40% of the data in the model building database can be used as the test dataset. The modeling platform is connected to build a PCB CNC model. The modeling platform is the tool used to build the model, such as MATLAB. Further, the PCB CNC model is trained and tested based on the training dataset and the test dataset to obtain a PCB CNC model that meets the preset accuracy. The PCB CNC model building process is as follows: each set of training data in the training dataset is input into the PCB CNC model. The output of the PCB CNC model is adjusted using the corresponding PCB image data and sample cutting trajectory data. When the output of the PCB CNC model is consistent with the sample cutting trajectory data, the training of the current set ends. The PCB CNC model training is completed when all the training data in the training dataset has been trained.

[0072] To ensure the accuracy of the CNC model, it can be tested using a test dataset. A preset accuracy rate can be set according to the actual situation. When the test accuracy rate of the test dataset meets the preset accuracy rate, the model training is complete. For example, the test accuracy rate can be set to 80%. When the test accuracy rate of the test dataset meets 80%, the CNC model construction is complete. Otherwise, the model needs to be retrained.

[0073] Input the target PCB image data into the board splitting CNC model and output the cutting trajectory data to improve cutting efficiency and accuracy.

[0074] Step S800: Perform data conversion on the cutting trajectory data based on the XYZ Cartesian mechanical coordinate system to obtain mechanical cutting trajectory data.

[0075] Specifically, the cutting of the PCB depanel is based on the XYZ Cartesian mechanical coordinate system. The cutting trajectory data obtained in step S700 is based on the target PCB image data. It is necessary to convert the cutting trajectory data to the XYZ Cartesian mechanical coordinate system to obtain mechanical cutting trajectory data, and then perform the cutting of the PCB depanel.

[0076] Based on the above analysis, this disclosure provides a high-precision segmentation method for PCB depaneling machines based on intelligent control. In this embodiment, the target PCB is scanned in a cyclic array to obtain image acquisition data. The image acquisition data is then processed, and through dynamic stitching and fusion of the images, a sample image training set is obtained. This sample image training set is used as training samples to train a depaneling CNC model based on a convolutional neural network, resulting in a converged depaneling CNC model. The model outputs the features of the position to be cut and feature points, thereby determining the PCB cutting trajectory data and achieving the technical effect of improving cutting efficiency and accuracy.

[0077] Example 2

[0078] Based on the same inventive concept as the high-precision PCB depaneling method based on intelligent control in the foregoing embodiments, such as Figure 2 As shown, this application also provides a high-precision PCB depaneling system based on intelligent control. The system includes an image acquisition device and a motion module. The system includes:

[0079] Coordinate system construction module 11, which is used to construct an XYZ Cartesian mechanical coordinate system;

[0080] Image acquisition device deployment module 12 is used to obtain a preset position based on the XYZ Cartesian mechanical coordinate system, and to deploy the image acquisition device based on the preset position to obtain the image acquisition device deployment result.

[0081] PCB physical acquisition module 13, the PCB physical acquisition module 13 is used to acquire the target PCB physical object;

[0082] The target cyclic scanning array acquisition module 14 is used to obtain the target cyclic scanning array by programming the cyclic moving scanning array based on the target PCB physical object through the motion module.

[0083] The cyclic image acquisition module 15 is used to control the image acquisition device to perform cyclic image acquisition on the target PCB object based on the deployment result of the image acquisition device and the target cyclic scanning array, and obtain image acquisition data by controlling the image acquisition device through the motion module.

[0084] Image processing module 16, which is used to perform image processing on the image acquisition data to obtain target PCB image data;

[0085] PCB CNC model construction module 17 is used to construct a PCB CNC model, input the target PCB image data into the PCB CNC model, and obtain cutting trajectory data.

[0086] The data conversion module 18 is used to convert the cutting trajectory data based on the XYZ Cartesian mechanical coordinate system to obtain mechanical cutting trajectory data.

[0087] Furthermore, the system also includes:

[0088] The motion module includes multiple preset PCB parameters and multiple preset cyclic moving scan arrays, and the multiple preset PCB parameters and the multiple preset cyclic moving scan arrays have a corresponding relationship;

[0089] The parameter acquisition module is used to acquire parameters of the target PCB and obtain the target PCB parameters.

[0090] A correlation evaluation module is used to perform correlation evaluation based on the target PCB parameters and the multiple preset PCB parameters to obtain correlation evaluation results for multiple PCB parameters.

[0091] A preset cyclic moving scan array filtering module is used to filter the multiple preset cyclic moving scan arrays based on the correlation evaluation results of the multiple PCB parameters to obtain the target cyclic scan array.

[0092] Furthermore, the system also includes:

[0093] The optimal correlation evaluation result acquisition module is used to determine the optimal correlation evaluation result based on the correlation evaluation results of the multiple PCB parameters.

[0094] A preset cyclic moving scan array matching module is used to match the multiple preset cyclic moving scan arrays based on the optimal correlation evaluation results to obtain the target cyclic scan array.

[0095] Furthermore, the system also includes:

[0096] An image quality assessment module is used to perform image quality assessment based on the image acquisition data and obtain image quality assessment coefficients.

[0097] An image quality constraint acquisition module is used to obtain image quality constraints.

[0098] An image quality evaluation coefficient determination module is used to determine whether the image quality evaluation coefficient satisfies the image quality constraint condition.

[0099] An optimized image acquisition instruction generation module is used to generate an optimized image acquisition instruction if the image quality evaluation coefficient does not meet the image quality constraint condition.

[0100] An image acquisition data update module is used to obtain optimized image acquisition data of the target PCB based on the optimized image acquisition command, and update the image acquisition data based on the optimized image acquisition data.

[0101] Furthermore, the system also includes:

[0102] A noise reduction processing module is used to perform noise reduction processing on the image acquisition data based on a median filtering algorithm to obtain denoised image acquisition data.

[0103] A stitching and fusion module is used to stitch and fuse the denoised image acquisition data to obtain an image fusion processing result;

[0104] An image enhancement module is used to enhance the image based on the image fusion processing result to obtain the target PCB image data.

[0105] Furthermore, the system also includes:

[0106] A feature location image data acquisition module is used to obtain feature location image data based on the image fusion processing result;

[0107] An image comparison module is used to compare the feature location image data with the denoised image acquisition data to obtain an image comparison result;

[0108] A stitching and fusion accuracy information acquisition module is used to obtain stitching and fusion accuracy information based on the image comparison results;

[0109] A stitching and blending adjustment module is used to adjust the image fusion processing result based on the stitching and blending accuracy information to obtain an adjusted image fusion result.

[0110] Furthermore, the system also includes:

[0111] A model building database acquisition module is used to acquire a model building database, wherein the model building database includes multiple sample PCB image data and multiple sample cutting trajectory data;

[0112] The model construction database random partitioning module is used to randomly partition the model construction database to obtain training datasets and test datasets;

[0113] The PCB CNC model construction module is used to connect to the modeling platform and construct the PCB CNC model.

[0114] The board-mounted CNC model acquisition module is used to train and test the board-mounted CNC model based on the training dataset and the test dataset to obtain the board-mounted CNC model that meets the preset accuracy.

[0115] The specific example of the high-precision PCB depaneling method based on intelligent control in Embodiment 1 described above is also applicable to the high-precision PCB depaneling system based on intelligent control in this embodiment. Through the foregoing detailed description of the high-precision PCB depaneling method based on intelligent control, those skilled in the art can clearly understand the high-precision PCB depaneling system based on intelligent control in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here. As for the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant details can be found in the method section.

[0116] Example 3

[0117] Figure 3 This is a schematic diagram based on the third embodiment of the present disclosure, as shown below. Figure 3 As shown, the electronic device 800 in this disclosure may include a processor 801 and a memory 802.

[0118] Memory 802 is used to store programs. Memory 802 may include volatile memory, such as random-access memory (RAM), such as static random-access memory (SRAM), double data rate synchronous dynamic random-access memory (DDR SDRAM), etc.; memory may also include non-volatile memory, such as flash memory. Memory 802 is used to store computer programs (such as application programs, functional modules, etc. that implement the above methods), computer instructions, etc. The computer programs, computer instructions, etc., can be partitioned and stored in one or more memories 802. Furthermore, the computer programs, computer instructions, data, etc., can be accessed by processor 801.

[0119] The aforementioned computer programs and instructions can be stored in one or more partitions of memory 802. Furthermore, the aforementioned computer programs and instructions can be invoked by processor 801.

[0120] The processor 801 is configured to execute the computer program stored in the memory 802 to implement the various steps in the methods described in the above embodiments.

[0121] For details, please refer to the relevant descriptions in the preceding method embodiments.

[0122] The processor 801 and the memory 802 can be independent structures or integrated structures. When the processor 801 and the memory 802 are independent structures, the memory 802 and the processor 801 can be coupled together via bus 803.

[0123] The electronic device in this embodiment can execute the technical solution in the above method. Its specific implementation process and technical principle are the same, and will not be repeated here.

[0124] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0125] According to embodiments of this disclosure, this disclosure also provides a computer program product comprising: a computer program stored in a readable storage medium, at least one processor of an electronic device being able to read the computer program from the readable storage medium, and the at least one processor executing the computer program causing the electronic device to perform the scheme provided in any of the above embodiments.

[0126] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders.

[0127] This document does not impose any restrictions as long as the desired results of the disclosed technical solution can be achieved.

[0128] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A high-precision PCB depaneling method based on intelligent control, characterized in that, The method is applied to a high-precision PCB depaneling system based on intelligent control. The system includes an image acquisition device and a motion module. The method includes: Establish an XYZ Cartesian machine coordinate system; Based on the XYZ Cartesian machine coordinate system, a preset position is obtained, and the image acquisition device is deployed based on the preset position to obtain the image acquisition device deployment result. Obtain the physical target PCB; Based on the target PCB physical object, the target cyclic scanning array is obtained by programming the cyclic moving scanning array through the motion module; Based on the deployment results of the image acquisition device and the target cyclic scanning array, the motion module controls the image acquisition device to perform cyclic image acquisition on the target PCB to obtain image acquisition data. The acquired image data is processed to obtain the target PCB image data; Construct a PCB splitting CNC model, and input the target PCB image data into the PCB splitting CNC model to obtain cutting trajectory data; Based on the XYZ Cartesian mechanical coordinate system, the cutting trajectory data is transformed to obtain mechanical cutting trajectory data; Specifically, based on the target PCB physical object, the motion module is used to program a cyclic scanning array to obtain the target cyclic scanning array, including: The motion module includes multiple preset PCB parameters and multiple preset cyclic moving scan arrays, and the multiple preset PCB parameters and the multiple preset cyclic moving scan arrays have a corresponding relationship; The parameters of the target PCB are obtained by collecting parameters from the physical target PCB. Based on the target PCB parameters and the multiple preset PCB parameters, a correlation evaluation is performed to obtain the correlation evaluation results of multiple PCB parameters; Based on the correlation evaluation results of the multiple PCB parameters, the multiple preset cyclic moving scan arrays are filtered to obtain the target cyclic scan array; After obtaining the image acquisition data, the process also includes: Based on the image acquisition data, image quality assessment is performed to obtain image quality assessment coefficients; Obtain image quality constraints; Determine whether the image quality evaluation coefficients meet the image quality constraints. If the image quality evaluation coefficient does not meet the image quality constraint, an optimized image acquisition instruction is generated. Based on the optimized image acquisition command, optimized image acquisition data of the target PCB is obtained, and the image acquisition data is updated based on the optimized image acquisition data.

2. The method as described in claim 1, characterized in that, The method further includes filtering the multiple preset cyclic moving scan arrays based on the correlation evaluation results of the multiple PCB parameters to obtain the target cyclic scan array. Based on the correlation evaluation results of the multiple PCB parameters, the optimal correlation evaluation result is determined; Based on the optimal correlation evaluation results, the multiple preset cyclic moving scan arrays are matched to obtain the target cyclic scan array.

3. The method as described in claim 1, characterized in that, The method further includes performing image processing on the acquired image data to obtain target PCB image data, and performing image processing on the acquired image data. The image acquisition data is denoised using a median filtering algorithm to obtain denoised image acquisition data. Based on the denoised image acquisition data, the images are stitched and fused to obtain the image fusion processing result; Image enhancement is performed based on the image fusion processing results to obtain the target PCB image data.

4. The method as described in claim 3, characterized in that, After obtaining the image fusion processing result, the method further includes: Based on the image fusion processing results, feature location image data is obtained; The image data at the feature location is compared with the denoised image acquisition data to obtain the image comparison result; Based on the image comparison results, the stitching and fusion accuracy information is obtained; Based on the stitching and fusion accuracy information, the image fusion processing result is stitched and fused to adjust, thereby obtaining the adjusted image fusion result.

5. The method as described in claim 1, characterized in that, The method for constructing a CNC model for sheet metal parts further includes: A model building database is obtained, wherein the model building database includes multiple sample PCB image data and multiple sample cutting trajectory data; The database for constructing the model is randomly divided to obtain training and test datasets. Connect to the modeling platform and construct the CNC model of the panel; The CNC model of the board is trained and tested based on the training dataset and the test dataset to obtain the CNC model of the board that meets the preset accuracy.

6. A high-precision PCB depaneling system based on intelligent control, characterized in that, The system is used to perform the method according to any one of claims 1 to 5, the system comprising an image acquisition device and a motion module, the system comprising: A coordinate system construction module, which is used to construct an XYZ Cartesian mechanical coordinate system; An image acquisition device deployment module is used to obtain a preset position based on the XYZ Cartesian mechanical coordinate system, and to deploy the image acquisition device based on the preset position to obtain the image acquisition device deployment result. A PCB physical acquisition module, which is used to acquire the target PCB physical object; A target cyclic scanning array acquisition module is used to obtain a target cyclic scanning array based on the target PCB physical object by programming a cyclic moving scanning array through the motion module. A cyclic image acquisition module is used to control the image acquisition device to perform cyclic image acquisition on the target PCB object based on the deployment result of the image acquisition device and the target cyclic scanning array, and obtain image acquisition data by controlling the image acquisition device through the motion module. The image processing module is used to perform image processing on the image acquisition data to obtain target PCB image data; The PCB CNC model construction module is used to construct a PCB CNC model, and inputs the target PCB image data into the PCB CNC model to obtain cutting trajectory data. The data conversion module is used to convert the cutting trajectory data based on the XYZ Cartesian mechanical coordinate system to obtain mechanical cutting trajectory data.

7. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.

Images