Bulk material placement machine chip positioning method

Abstract

The present invention provides a patch positioning method for a bulk patch machine, including: obtaining characteristic information of a circuit board, determining a corresponding image processing model based on the characteristic information of the circuit board; starting a camera on the bulk patch machine to capture an image of the circuit board, analyzing the image based on the image processing model, and identifying the actual position relationship between a reference mark point on the circuit board and a position to be patched; obtaining an ideal position relationship between the reference mark point and the position to be patched; correcting the motion path of a patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship; and controlling the patch head to be positioned to the position to be patched based on the corrected motion path. In the present invention, the actual position relationship between the reference mark point and the position to be patched is automatically identified by analyzing the image of the circuit board, and then the motion path of the patch head is dynamically adjusted to ensure that components can be accurately patched.

Description

Technical Field

[0001] The invention relates to the technical field of image processing, and in particular to a patch positioning method of a bulk material patch machine. Background Art

[0002] In the electronics manufacturing industry, bulk chip mounters are one of the key devices for achieving high-efficiency and high-precision electronic component assembly. They are mainly responsible for accurately placing electronic components of various sizes on the predetermined positions of the circuit board (PCB).

[0003] At present, various factors in the patch production process, such as circuit board deformation, positioning error, machine vibration, etc., may cause deviations in the patch position, thereby affecting the patch accuracy of the final product.

[0004] The traditional method mainly adopts manual adjustment of the patch position, but this method is time-consuming, labor-intensive and inefficient. Summary of the invention

[0005] The main purpose of the present invention is to provide a patch positioning method for a bulk patch machine, aiming to overcome the defects of current patch accuracy and low efficiency.

[0006] To achieve the above object, the present invention provides a patch positioning method for a bulk patch machine, comprising the following steps:

[0007] Acquire characteristic information of the circuit board, and determine a corresponding image processing model based on the characteristic information of the circuit board;

[0008] Starting a camera on the bulk placement machine to capture an image of the circuit board, analyzing the image based on the image processing model, and identifying the actual positional relationship between the reference mark point on the circuit board and the position to be placed;

[0009] Acquire an ideal position relationship between the reference mark point and the position to be mounted;

[0010] According to the actual position relationship and the ideal position relationship, the movement path of the patch head of the bulk patch machine is corrected; based on the corrected movement path, the patch head is controlled to be positioned to the position to be patched.

[0011] Furthermore, the correcting of the motion path of the patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship includes:

[0012] Calculating a deviation between the actual position relationship and the ideal position relationship, and obtaining a deviation between the position to be patched and the ideal patch position according to the deviation;

[0013] Obtaining an X-axis deviation value and a Y-axis deviation value from the deviation between the position to be mounted and the ideal position of the patch;

[0014] Based on the X-axis deviation value and the Y-axis deviation value, the motion path of the placement head of the bulk placement machine is corrected.

[0015] Furthermore, the step of obtaining an ideal position relationship between the reference mark point and the position to be mounted includes:

[0016] Obtaining model information of the circuit board and the current process flow of the circuit board;

[0017] According to the model information, a corresponding target parameter data table is matched in a database; wherein the database stores a mapping relationship between the model information and the parameter data table; and the parameter data table records the process parameters of each process flow;

[0018] In the target parameter data table, searching for the target process parameter corresponding to the current process flow;

[0019] The ideal position relationship between the reference mark point and the position to be mounted is obtained from the target process parameters.

[0020] Furthermore, the acquiring characteristic information of the circuit board and determining the corresponding image processing model based on the characteristic information of the circuit board includes:

[0021] Acquire an overall image of the circuit board;

[0022] Identifying circuit layout information and key device information from the overall image;

[0023] Obtaining specific type information of the circuit board based on the circuit layout information and key component information;

[0024] Based on the mapping relationship between the type information and the recognition model stored in the database, the recognition model corresponding to the specific type information is obtained as the image processing model.

[0025] Furthermore, the acquiring characteristic information of the circuit board and determining the corresponding image processing model based on the characteristic information of the circuit board includes:

[0026] Scanning a barcode in a designated area of ​​the circuit board, and parsing the barcode to obtain stored character combination information; wherein the character combination information includes a combination of multiple English characters and numeric characters;

[0027] Extracting all the numeric characters in the character combination information in sequence, and combining them in sequence to obtain a numeric character combination;

[0028] Obtaining manufacturer information of a manufacturer that performs patch processing on the circuit board;

[0029] Based on the manufacturer information, correcting the preset conversion table to obtain a corrected conversion table;

[0030] Based on the correction conversion table, convert the digital character combination to obtain a conversion character combination;

[0031] Based on the mapping relationship between the character combination and the recognition model, the recognition model of the converted character combination mapping is obtained as the image processing model.

[0032] Further, based on the manufacturer information, correcting the preset conversion table to obtain a corrected conversion table includes:

[0033] Obtain a preset conversion table; wherein the conversion table includes a number column and a conversion column, and the numeric characters in the number column and the English characters in the conversion column have a one-to-one conversion relationship;

[0034] Extract all English characters from the manufacturer information to obtain the manufacturer English information, and obtain the total number of English characters in the manufacturer English information;

[0035] Deduplication of the manufacturer's English information is performed to obtain deduplication manufacturer information;

[0036] Extracting target characters identical to the English characters in the deduplicated vendor information from the conversion column of the conversion table, and combining the target characters according to their original order in the conversion table to obtain a target character combination;

[0037] Insert each English character in the target character combination in sequence starting from the mth position of the conversion column, and shift other English characters in the conversion column backward in sequence to fill the conversion column completely, thereby obtaining the corrected conversion table; wherein m is a number corresponding to the total number of English characters in the manufacturer's English information.

[0038] Further, based on the manufacturer information, correcting the preset conversion table to obtain a corrected conversion table includes:

[0039] Create a blank data table with three rows and three columns;

[0040] Add the first nine characters in the manufacturer information to the blank data table one by one in sequence to obtain a character data table;

[0041] In the character data table, respectively detecting whether the characters in three rows and three columns are pure numeric characters or pure English characters;

[0042] Get the total number of columns x and the total number of rows y that are pure numeric characters or pure English characters in each row and column;

[0043] Obtain a preset conversion table; wherein the conversion table includes a number column and a conversion column, and the numeric characters in the number column and the English characters in the conversion column have a one-to-one conversion relationship;

[0044] A first prefix is ​​added before the English characters in the conversion column corresponding to all odd numbers in the number column, and a first suffix is ​​added after the English characters in the conversion column corresponding to all even numbers in the number column; wherein the first prefix is ​​the number corresponding to x, and the first suffix is ​​the number corresponding to y.

[0045] Furthermore, the analyzing the image based on the image processing model to identify the actual positional relationship between the reference marking point on the circuit board and the position to be mounted includes:

[0046] Inputting the image into the image processing model; wherein the image processing model includes a pre-trained key point recognition sub-model and a position relationship recognition sub-model;

[0047] Performing preliminary recognition on the image to obtain a preliminary recognition result, and adjusting model parameters of a key point recognition sub-model and a position relationship recognition sub-model based on the preliminary recognition result;

[0048] Analyzing the image based on the adjusted key point recognition sub-model to obtain reference marking points and positions to be mounted on the circuit board in the image, and marking the reference marking points and positions to be mounted on the image to obtain a marked image;

[0049] The actual position relationship analysis of the marked image is performed based on the adjusted position relationship recognition sub-model to obtain the actual position relationship between the reference marking point on the circuit board in the marked image and the position to be mounted.

[0050] The present invention also provides a patch positioning device for a bulk patch machine, comprising:

[0051] A first acquisition unit, which acquires characteristic information of a circuit board and determines a corresponding image processing model based on the characteristic information of the circuit board;

[0052] An identification unit is used to start a camera on the bulk placement machine to collect an image of the circuit board, analyze the image based on the image processing model, and identify the actual position relationship between the reference mark point on the circuit board and the position to be placed;

[0053] A second acquisition unit, used for acquiring an ideal position relationship between the reference mark point and the position to be mounted;

[0054] The positioning unit is used to correct the movement path of the patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship; based on the corrected movement path, the patch head is controlled to be positioned to the position to be patched.

[0055] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0056] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0057] The patch positioning method of the bulk patch machine provided by the present invention includes: obtaining characteristic information of a circuit board, and determining a corresponding image processing model based on the characteristic information of the circuit board; starting a camera on the bulk patch machine to capture an image of the circuit board, analyzing the image based on the image processing model, and identifying the actual position relationship between the reference mark point on the circuit board and the position to be patched; obtaining the ideal position relationship between the reference mark point and the position to be patched; correcting the motion path of the patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship; and controlling the patch head to be positioned to the position to be patched based on the corrected motion path. In the present invention, the actual position relationship between the reference mark point and the position to be patched is automatically identified by analyzing the image of the circuit board, and then the motion path of the patch head is dynamically adjusted to ensure that the component can be accurately patched to the target position. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the steps of a patch positioning method of a bulk patch machine in one embodiment of the present invention;

[0059] Figure 2 It is a structural block diagram of a patch positioning device of a bulk material patch machine in one embodiment of the present invention;

[0060] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0061] The implementation, functional features and advantages of the present invention will be further described in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0063] Reference Figure 1 In one embodiment of the present invention, a patch positioning method for a bulk patch placement machine is provided, comprising the following steps:

[0064] Step S1, acquiring characteristic information of a circuit board, and determining a corresponding image processing model based on the characteristic information of the circuit board;

[0065] Step S2, starting a camera on the bulk placement machine to capture an image of the circuit board, analyzing the image based on the image processing model, and identifying the actual positional relationship between the reference mark point on the circuit board and the position to be placed;

[0066] Step S3, obtaining an ideal position relationship between the reference mark point and the position to be mounted;

[0067] Step S4, correcting the motion path of the placement head of the bulk placement machine according to the actual position relationship and the ideal position relationship; and controlling the placement head to be positioned to the position to be placed based on the corrected motion path.

[0068] In this embodiment, it specifically includes:

[0069] Characteristic information of circuit boards: By analyzing the physical characteristics and structural layout of the circuit boards, such as size, distribution of marking points, material properties, etc., this information is crucial to building an efficient and accurate image processing model. Different image processing models should be used for different circuit boards to improve the model processing effect. Based on the extracted feature information, select or customize an image processing model. The above model can be a neural network based on deep learning, a support vector machine (SVM), or other machine learning algorithms, which aims to optimize the recognition accuracy for specific circuit board designs. In this embodiment, different circuit boards are adapted to different image processing models, which is beneficial to improving the adaptability and pertinence of the image processing model, and facilitates better subsequent image analysis.

[0070] Image acquisition: Bulk SMT machines are pre-installed with cameras, and use cameras or other imaging devices on bulk SMT machines to capture images of circuit boards. These devices can clearly record tiny details, including the texture, color differences, and preset marking points on the surface of the circuit board.

[0071] Image preprocessing: The collected images will go through a series of preprocessing steps, such as denoising, contrast enhancement, grayscale or binarization, to remove unnecessary interference information and highlight the features of the marker points and the area to be patched.

[0072] Feature recognition: Use image recognition algorithms (such as edge detection, template matching, feature point extraction, etc.) to automatically identify reference markers (such as positioning holes, special graphic marks, etc.) and the positions to be mounted from the pre-processed images. The above markers are the basis for accurate mounting, ensuring that the mounting position is consistent with the design intention. Figure 1 To.

[0073] Position calculation: Apply the selected image processing model to accurately calculate the spatial position of the identified reference markers and the positions to be mounted, and obtain the actual relative position between them. This step is crucial to understanding the actual layout of the components to be mounted on the current circuit board.

[0074] Deviation calculation: Define the ideal position relationship in advance or extract it from the design file, that is, the theoretically optimal patch coordinates. By comparing the actual measured position with the ideal position, calculate the deviation between the two.

[0075] Path planning algorithm: Based on the identified position deviation, the path planning algorithm is used to adjust the movement trajectory of the patch head. This involves fine-tuning the original path, bending compensation, or complete replanning to ensure that the patch head can accurately reach the corrected patch point.

[0076] Real-time control: This correction process is carried out in real time, which means that the placement machine can respond quickly based on the latest visual feedback before each placement, reflecting a high degree of flexibility and adaptability.

[0077] Precise positioning: According to the corrected motion path, the servo system accurately controls the movement of the patch head to ensure that it can be positioned at the target patch position with extremely high accuracy. At this time, the patch head picks up the component from the feeder and releases it to the circuit board to complete the patch action.

[0078] In summary, the core of the above technical solution is to combine advanced visual recognition technology with precise motion control technology. Through intelligent analysis and real-time adjustment, it overcomes the positioning accuracy limitations faced by traditional placement machines, especially when dealing with complex circuit board designs and tiny component placements. This not only significantly improves production efficiency and reduces defective rates, but also brings higher flexibility and intelligence to the electronics manufacturing industry.

[0079] In one embodiment, the correcting the motion path of the placement head of the bulk placement machine according to the actual position relationship and the ideal position relationship includes:

[0080] Calculating a deviation between the actual position relationship and the ideal position relationship, and obtaining a deviation between the position to be patched and the ideal patch position according to the deviation;

[0081] Obtaining an X-axis deviation value and a Y-axis deviation value from the deviation between the position to be mounted and the ideal position of the patch;

[0082] Based on the X-axis deviation value and the Y-axis deviation value, the motion path of the placement head of the bulk placement machine is corrected.

[0083] In this embodiment, it specifically includes:

[0084] Deviation value determination: First, calculate the deviation between the actual position relationship (i.e., the patch position obtained by visual recognition) and the ideal position relationship (design requirements or theoretically correct position). This deviation value comprehensively reflects the horizontal (X-axis) and vertical (Y-axis) offsets of the patch position, as well as possible rotation angle deviations (although rotation deviations are not involved in this embodiment, they can also be comprehensively considered in other embodiments, and will not be described in detail in this embodiment).

[0085] Two-dimensional deviation decomposition: For ease of understanding and adjustment, the overall deviation is further decomposed into specific values ​​in the X-axis and Y-axis directions. The X-axis deviation value represents the offset along the width of the circuit board, while the Y-axis deviation value represents the offset along the length. These two values ​​directly indicate how much correction the patch head needs to make in each of the two dimensions to achieve the ideal position.

[0086] Path adjustment strategy: Based on the deviation values ​​of the X-axis and Y-axis, the motion path of the patch head is corrected. This includes directly adding the corresponding offset to the original motion command, or in complex scenarios, replanning an optimal path from the current position to the target position through an iterative optimization algorithm. Path correction not only considers the deviation correction of linear movement, but may also cover speed control, acceleration adjustment, etc. to ensure a smooth and precise motion trajectory.

[0087] Servo system response: The corrected motion path command is transmitted to the servo control system of the bulk placement machine, which can respond quickly and accurately adjust the movement of the placement head. The servo motor adjusts its rotation angle and speed according to the new path command to ensure that the placement head reaches each placement point accurately according to the corrected path.

[0088] In summary, through the above steps, this technical solution realizes real-time and precise adjustment of the movement path of the patch machine, effectively solving the problem of deviation between the actual patch position and the ideal position caused by various factors. Such a mechanism greatly improves the positioning accuracy of the patch, ensuring that even the patch of high-density and tiny components can meet the design requirements, improving the overall production efficiency and product quality.

[0089] In one embodiment, the step of obtaining an ideal position relationship between the reference mark point and the position to be mounted includes:

[0090] Obtaining model information of the circuit board and the current process flow of the circuit board;

[0091] According to the model information, a corresponding target parameter data table is matched in a database; wherein the database stores a mapping relationship between the model information and the parameter data table; and the parameter data table records the process parameters of each process flow;

[0092] In the target parameter data table, searching for the target process parameter corresponding to the current process flow;

[0093] The ideal position relationship between the reference mark point and the position to be mounted is obtained from the target process parameters.

[0094] In this embodiment, it specifically includes:

[0095] Model identification and process confirmation: First, it is necessary to identify the model information of the circuit board and determine the process the circuit board is currently undergoing. The model information is the unique identifier of the circuit board design and specifications, while the process refers to the specific processing stage of the circuit board on the production line, such as component placement, welding, etc.

[0096] Database query and mapping: After obtaining the PCB model information, it will be queried in the pre-established database. The database maintains a mapping relationship, which associates different PCB models with their corresponding parameter data tables. Each parameter data table is designed for a specific model of PCB and contains all the necessary parameters under different process flows.

[0097] Parameter table location: According to the identified circuit board model information, find the matching target parameter data table from the database. This data table records the detailed process parameters applicable to this model of circuit board at each process stage, including but not limited to patch position, temperature setting, time control, etc.

[0098] Ideal position data extraction: Then, according to the current process flow, find and extract the corresponding ideal position relationship data in the target parameter data table. These data directly define the ideal relative position and direction between the reference mark point and each position to be mounted in the process flow, which is the basis for the precision control of the mounting.

[0099] In summary, this technical solution ensures that the placement machine can quickly and accurately obtain the required placement position information according to the specific model and production stage of the circuit board through an intelligent data management and retrieval mechanism. This process not only greatly simplifies the work of production preparation and parameter configuration, but also greatly improves production efficiency and placement accuracy, and is an important part of realizing intelligent and efficient electronic manufacturing. By utilizing database storage and fast query technology, the solution flexibly responds to diverse product models and complex production process requirements, demonstrating the advantages of modern intelligent manufacturing technology.

[0100] In one embodiment, the acquiring characteristic information of the circuit board and determining a corresponding image processing model based on the characteristic information of the circuit board includes:

[0101] Acquire an overall image of the circuit board;

[0102] Identifying circuit layout information and key device information from the overall image;

[0103] Obtaining specific type information of the circuit board based on the circuit layout information and key component information;

[0104] Based on the mapping relationship between the type information and the recognition model stored in the database, the recognition model corresponding to the specific type information is obtained as the image processing model.

[0105] In this embodiment, it specifically includes:

[0106] Image acquisition: First, a high-resolution camera or scanning device is used to capture a complete image of the circuit board. This step provides comprehensive visual data for subsequent analysis and is the basis for identifying circuit board features.

[0107] Circuit layout and key component identification: Through image processing and computer vision technology, the layout features of the circuit board are identified from the overall image, such as wire paths, via distribution, and the location and type of key components (such as large integrated circuits, special packaging components, etc.). This information is an intuitive reflection of the characteristics of the circuit board and is crucial for distinguishing different types of circuit boards.

[0108] PCB type determination: Based on the identified circuit layout and key component information, further analysis and determination of the specific type of PCB is performed. Different types of PCBs have different patch requirements and layout rules. Therefore, accurate identification of PCB types is a prerequisite for customized image processing models.

[0109] Database mapping query: A database is maintained in advance, which stores the mapping relationship between various circuit board type information and the corresponding recommended image processing models. Each type of circuit board has its most suitable recognition model, which is designed for specific layouts, component arrangements or marking points, and can perform image recognition and analysis more efficiently and accurately.

[0110] Model selection: According to the specific type information of the circuit board, the corresponding image processing model is searched and selected in the database. This model will be used for subsequent image analysis to identify reference markers and locations to be mounted, ensuring the accuracy and efficiency of recognition.

[0111] In summary, through the above steps, this technical solution realizes personalized recognition and processing of different circuit boards, ensuring the pertinence and efficiency of image recognition. From obtaining the global image of the circuit board to identifying key information, and then matching the most appropriate image processing model according to the type of circuit board, this series of processes demonstrates a high degree of intelligence and adaptability. In this way, no matter what kind of circuit board design is faced, the placement machine can quickly and accurately locate and place components through precise image processing models, thereby improving production efficiency and placement accuracy.

[0112] In one embodiment, the analyzing the image based on the image processing model to identify the actual positional relationship between the reference mark point on the circuit board and the position to be mounted includes:

[0113] Inputting the image into the image processing model; wherein the image processing model includes a pre-trained key point recognition sub-model and a position relationship recognition sub-model;

[0114] Performing preliminary recognition on the image to obtain a preliminary recognition result, and adjusting model parameters of a key point recognition sub-model and a position relationship recognition sub-model based on the preliminary recognition result;

[0115] Analyzing the image based on the adjusted key point recognition sub-model to obtain reference marking points and positions to be mounted on the circuit board in the image, and marking the reference marking points and positions to be mounted on the image to obtain a marked image;

[0116] The actual position relationship analysis of the marked image is performed based on the adjusted position relationship recognition sub-model to obtain the actual position relationship between the reference marking point on the circuit board in the marked image and the position to be mounted.

[0117] In this embodiment, it specifically includes:

[0118] Image input and preliminary recognition: First, the circuit board image to be analyzed is input into a pre-trained image processing model. This model contains two sub-models: key point recognition sub-model and position relationship recognition sub-model. In the preliminary recognition stage, the model will perform preliminary analysis on the image according to the preset algorithm and output the preliminary recognition result.

[0119] Model parameter adjustment: Based on the preliminary recognition results, the accuracy and reliability of the model prediction will be evaluated. If it is found that there is a large deviation between the prediction result and the actual situation, the parameters of the key point recognition sub-model and the position relationship recognition sub-model will be automatically adjusted to optimize the model performance and improve the accuracy of subsequent recognition.

[0120] Key point recognition and marking: After parameter adjustment, the key point recognition sub-model will analyze the image more accurately and identify the reference marking points and the positions of the chips to be mounted on the circuit board. These position information will be marked on the image to generate a marked image, which is convenient for subsequent position relationship analysis.

[0121] Actual position relationship analysis: The position relationship recognition sub-model further analyzes the actual position relationship between the reference mark points on the circuit board and the position to be mounted based on the mark image. This step is crucial to ensure that the placement machine can accurately place electronic components in the correct position.

[0122] In this embodiment, the whole process is dynamic, and as the model continues to process more image data, its recognition accuracy will gradually improve. In addition, by continuously collecting and analyzing feedback from actual production, the model parameters can be further optimized to improve the stability and reliability of the overall system.

[0123] The application of the above technical solutions is not limited to the field of circuit board manufacturing, but can also be extended to other scenarios that require high-precision image recognition and positioning, such as industrial automation, robot vision, etc. Through deep learning and computer vision technology, automated and intelligent image processing is achieved, greatly improving production efficiency and product quality.

[0124] In this embodiment, preliminary recognition refers to the first analysis of the input image by the model to identify key features, such as reference markers on the circuit board and the locations of the patches to be placed. This stage usually uses a pre-trained model, which has learned some basic patterns and features from a large amount of historical image data. The results of preliminary recognition may not be perfect, but they can provide a basic framework to guide subsequent more detailed analysis.

[0125] The initial identification process includes the following steps:

[0126] Feature extraction: The model extracts key features from the image, such as edges, textures, shapes, etc.

[0127] Feature mapping: Map the extracted features into the feature space of the model for classification or recognition.

[0128] Initial prediction: Based on the mapped features, the model attempts to identify the target objects or regions in the image, such as markers and patch locations on a circuit board.

[0129] Model parameter adjustment:

[0130] After the initial identification, the model's prediction results will be evaluated to see how well they match the actual situation. If the model's prediction error is large, the model's parameters need to be adjusted to improve the model's performance. Parameter adjustment is usually achieved in the following ways:

[0131] Backpropagation algorithm: This is the most commonly used parameter update mechanism in deep learning. It calculates the gradient of the loss function (a function that measures the difference between the model's predicted value and the true value) relative to each weight, and then updates the weights at a certain learning rate to minimize the loss function.

[0132] Optimization algorithms: such as stochastic gradient descent (SGD), Adam, RMSprop, etc., which can help the model converge to the optimal solution faster.

[0133] Data enhancement: To make the model more robust, the training data can be enhanced by rotation, flipping, scaling, etc. to increase the generalization ability of the model.

[0134] Hyperparameter adjustment: including learning rate, batch size, number of network layers, etc. The choice of these hyperparameters will also affect the performance of the model.

[0135] The process of adjusting model parameters is an iterative cycle, and each iteration will further optimize the parameters based on the new model prediction results until the model reaches satisfactory performance indicators. In this way, the model can gradually learn and adapt to the needs of specific tasks, and ultimately achieve high-precision key point recognition and position relationship analysis.

[0136] In this embodiment, preliminary recognition and model parameter adjustment are the core links of deep learning model optimization, ensuring that the model can gradually evolve from the initial rough estimate to a highly accurate prediction. The above process reflects the powerful adaptive ability and optimization potential of deep learning, which is the key to achieving complex image recognition tasks.

[0137] In one embodiment, the step of acquiring characteristic information of a circuit board and determining a corresponding image processing model based on the characteristic information of the circuit board includes:

[0138] Scanning a barcode in a designated area of ​​the circuit board, and parsing the barcode to obtain stored character combination information; wherein the character combination information includes a combination of multiple English characters and numeric characters;

[0139] Extracting all the numeric characters in the character combination information in sequence, and combining them in sequence to obtain a numeric character combination;

[0140] Obtaining manufacturer information of a manufacturer that performs patch processing on the circuit board;

[0141] Based on the manufacturer information, correcting the preset conversion table to obtain a corrected conversion table;

[0142] Based on the correction conversion table, convert the digital character combination to obtain a conversion character combination;

[0143] Based on the mapping relationship between the character combination and the recognition model, the recognition model of the converted character combination mapping is obtained as the image processing model.

[0144] In this embodiment, it specifically includes:

[0145] Barcode recognition: First, use a barcode scanner or a camera with barcode recognition function to scan the barcode in the specified area on the circuit board. The barcode stores a string of information containing English characters and numeric characters, which is a coded representation of the identity and specific attributes of the circuit board.

[0146] Character extraction and combination: From the parsed character combination information, only all numeric characters are extracted and combined in sequence to form a new numeric character combination. This processing step simplifies the information processing flow and focuses on those numeric codes that are directly related to the selection of the recognition model.

[0147] Manufacturer information identification: Get information about the manufacturer responsible for mounting the circuit board. Different manufacturers have their own standards or preferences, which affects how to interpret the information in the barcode.

[0148] Conversion table correction: Based on the obtained manufacturer information, a preset conversion table is appropriately adjusted and corrected. This conversion table is a mapping rule used to convert digital character combinations into character combinations that can be directly recognized by the internal system. The correction process ensures the accuracy and compatibility of model selection, taking into account the differentiated requirements of different manufacturers.

[0149] Convert character combination: Use the corrected conversion table to convert the previously extracted digital character combination to obtain a new converted character combination. This converted character combination is the direct basis for selecting the image processing model.

[0150] Model mapping and selection: Finally, according to the mapping relationship between the converted character combination and the pre-defined character combination and recognition model, the corresponding image processing model is found. This model is specially optimized for the specific type and manufacturer requirements of the circuit board, and can perform visual recognition and positioning more accurately. The above process is beneficial to improve the security and uniqueness of the data processing process and avoid data leakage.

[0151] In summary, this technical solution cleverly utilizes the barcode information on the circuit board, and through a series of logical processing and conversion steps, efficiently and accurately determines the image processing model suitable for a specific circuit board and processing environment. This process not only enhances the adaptability to different production conditions, but also achieves more personalized model selection by considering manufacturer information, which helps to improve the overall recognition accuracy and production efficiency of the placement machine.

[0152] In one embodiment, the step of correcting a preset conversion table based on the manufacturer information to obtain a corrected conversion table includes:

[0153] Obtain a preset conversion table; wherein the conversion table includes a number column and a conversion column, and the numeric characters in the number column and the English characters in the conversion column have a one-to-one conversion relationship;

[0154] Extract all English characters from the manufacturer information to obtain the manufacturer English information, and obtain the total number of English characters in the manufacturer English information;

[0155] Deduplication of the manufacturer's English information is performed to obtain deduplication manufacturer information;

[0156] Extracting target characters identical to the English characters in the deduplicated vendor information from the conversion column of the conversion table, and combining the target characters according to their original order in the conversion table to obtain a target character combination;

[0157] Insert each English character in the target character combination in sequence starting from the mth position of the conversion column, and shift other English characters in the conversion column backward in sequence to fill the conversion column completely, thereby obtaining the corrected conversion table; wherein m is a number corresponding to the total number of English characters in the manufacturer's English information.

[0158] In this embodiment, it specifically includes:

[0159] Preset conversion table: First, obtain a preset conversion table. The conversion table is preset and contains two columns, the first column is a number column, and the second column is a conversion column. In each row, there is a one-to-one correspondence between the numeric characters in the number column and the English characters in the conversion column, which is the original character conversion rule.

[0160] Extract English characters: Extract all English characters from the acquired manufacturer information, count the total number of these English characters, and use this as an important basis for subsequent operations.

[0161] De-duplication: De-duplication is performed on the extracted manufacturer's English information to ensure that each English character appears only once, and the manufacturer's English information after de-duplication is obtained. This operation is to simplify the subsequent conversion rules and avoid repeated operations.

[0162] Extract target characters: In the conversion column of the preset conversion table, find the English characters that are the same as those in the English information of the manufacturer after deduplication, keep their original order in the conversion table, and combine these characters together to form the target character combination.

[0163] Insertion and translation: Based on the total number of English characters in the deduplicated manufacturer information (denoted as m), starting from the mth position of the conversion column, the English characters in the target character combination are inserted into the conversion column in sequence. This means that the English characters starting from the mth position in the original conversion column need to be translated backwards in sequence to make room for the newly inserted characters while ensuring the integrity of the conversion column, thereby generating a corrected conversion table for a specific manufacturer. It is understandable that if there are English characters translated out of the end of the character column, the translated characters will be translated to the head of the character column to make the character column complete.

[0164] In summary, this technical solution uses a highly customized processing flow to dynamically adjust the character conversion rules according to the specific information of the circuit board processing manufacturer, and generates a correction conversion table that better matches the current production environment. This unique approach not only improves the flexibility and adaptability of the recognition system, but also improves the security of the data processing process. The uniqueness of the correction process of the conversion table can avoid being deciphered; at the same time, it ensures that the selection of the image processing model is more accurate, meets the personalized needs of different manufacturers, and further optimizes the chip positioning accuracy and efficiency of the placement machine. The entire process cleverly uses operations such as information extraction, deduplication, character insertion and translation, demonstrating a high degree of logic and practicality.

[0165] In one embodiment, the step of correcting a preset conversion table based on the manufacturer information to obtain a corrected conversion table includes:

[0166] Create a blank data table with three rows and three columns;

[0167] Add the first nine characters in the manufacturer information to the blank data table one by one in sequence to obtain a character data table;

[0168] In the character data table, respectively detecting whether the characters in three rows and three columns are pure numeric characters or pure English characters;

[0169] Get the total number of columns x and the total number of rows y that are pure numeric characters or pure English characters in each row and column;

[0170] Obtain a preset conversion table; wherein the conversion table includes a number column and a conversion column, and the numeric characters in the number column and the English characters in the conversion column have a one-to-one conversion relationship;

[0171] A first prefix is ​​added before the English characters in the conversion column corresponding to all odd numbers in the number column, and a first suffix is ​​added after the English characters in the conversion column corresponding to all even numbers in the number column; wherein the first prefix is ​​the number corresponding to x, and the first suffix is ​​the number corresponding to y.

[0172] In this embodiment, it specifically includes:

[0173] First, create a blank data table with three rows and three columns. This is a basic framework for storing and processing characters in manufacturer information.

[0174] Fill the first nine characters of the manufacturer information into this three-row, three-column data table in order. This is done so that the nature of these characters can be analyzed by row and column later, that is, whether they are pure numbers or pure English characters.

[0175] In this character data table, check the character combination of each row and column to determine whether they are all pure numeric characters or pure English characters. This step is to count the distribution of pure numeric characters or pure English characters.

[0176] Based on the above detection, the total number of pure numeric characters or pure English characters in all rows and columns is counted, marked as the total number of columns x and the total number of rows y respectively. x and y will be used to determine the values ​​of the prefix and suffix.

[0177] Next, a preset conversion table is obtained. This conversion table has two key parts: a number column and a conversion column, where each number in the number column corresponds to an English character in the conversion column.

[0178] Modify this preset conversion table, add x (the total number of columns of pure numbers or English characters) as a prefix to the conversion column English characters corresponding to all odd numbers in the digital column; and add y (the total number of rows of pure numbers or English characters) as a suffix to the conversion column English characters corresponding to all even numbers. This operation is intended to make targeted adjustments to the original conversion rules based on the characteristics of the manufacturer's information, in order to enhance the security and uniqueness of data processing and prevent it from being easily deciphered by a third party.

[0179] In summary, the core of the above technical solution is to customize and modify the preset character conversion rules by analyzing the character characteristics in the manufacturer information (especially the distribution of pure numbers or English characters in the data table). By adding specific prefixes and suffixes, the conversion table can better adapt to the needs of specific manufacturers, which helps to improve the security, uniqueness and accuracy of data processing, especially when processing input data in a specific format. The whole process reflects the operation of data structure, the identification of character types and the adaptive adjustment of rules based on statistical information.

[0180] Reference Figure 2 In another embodiment of the present invention, a patch positioning device for a bulk patch machine is provided, comprising:

[0181] A first acquisition unit, which acquires characteristic information of a circuit board and determines a corresponding image processing model based on the characteristic information of the circuit board;

[0182] An identification unit is used to start a camera on the bulk placement machine to collect an image of the circuit board, analyze the image based on the image processing model, and identify the actual position relationship between the reference mark point on the circuit board and the position to be placed;

[0183] A second acquisition unit, used for acquiring an ideal position relationship between the reference mark point and the position to be mounted;

[0184] The positioning unit is used to correct the movement path of the patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship; based on the corrected movement path, the patch head is controlled to be positioned to the position to be patched.

[0185] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the above method embodiment, which will not be repeated here.

[0186] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0187] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0188] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0189] In summary, the patch positioning method of the bulk patch machine provided in the embodiment of the present invention includes: obtaining characteristic information of the circuit board, and determining the corresponding image processing model based on the characteristic information of the circuit board; starting the camera on the bulk patch machine to capture the image of the circuit board, analyzing the image based on the image processing model, and identifying the actual position relationship between the reference mark point on the circuit board and the position to be patched; obtaining the ideal position relationship between the reference mark point and the position to be patched; correcting the motion path of the patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship; and controlling the patch head to be positioned to the position to be patched based on the corrected motion path. In the present invention, the actual position relationship between the reference mark point and the position to be patched is automatically identified by analyzing the image of the circuit board, and then the motion path of the patch head is dynamically adjusted to ensure that the component can be accurately patched to the target position.

[0190] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0191] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0192] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A chip positioning method for a bulk chip placement machine, characterized in that: The following steps are involved: Acquire characteristic information of the circuit board, and determine a corresponding image processing model based on the characteristic information of the circuit board; Starting a camera on the bulk placement machine to capture an image of the circuit board, analyzing the image based on the image processing model, and identifying the actual positional relationship between the reference mark point on the circuit board and the position to be placed; Acquire an ideal position relationship between the reference mark point and the position to be mounted; Correcting the motion path of a placement head of a bulk placement machine according to the actual position relationship and the ideal position relationship; Based on the corrected motion path, control the patch head to be positioned at the position to be patched; The step of acquiring the ideal position relationship between the reference mark point and the position to be mounted includes: Obtaining model information of the circuit board and the current process flow of the circuit board; According to the model information, a corresponding target parameter data table is matched in a database; wherein the database stores a mapping relationship between the model information and the parameter data table; and the parameter data table records the process parameters of each process flow; In the target parameter data table, searching for the target process parameter corresponding to the current process flow; Acquire the ideal position relationship between the reference mark point and the position to be mounted from the target process parameters; The analyzing the image based on the image processing model to identify the actual positional relationship between the reference marking point on the circuit board and the position of the patch to be mounted includes: Inputting the image into the image processing model; wherein the image processing model includes a pre-trained key point recognition sub-model and a position relationship recognition sub-model; Performing preliminary recognition on the image to obtain a preliminary recognition result, and adjusting model parameters of a key point recognition sub-model and a position relationship recognition sub-model based on the preliminary recognition result; Analyzing the image based on the adjusted key point recognition sub-model to obtain reference marking points and positions to be mounted on the circuit board in the image, and marking the reference marking points and positions to be mounted on the image to obtain a marked image; The actual position relationship analysis of the marked image is performed based on the adjusted position relationship recognition sub-model to obtain the actual position relationship between the reference marking point on the circuit board in the marked image and the position to be mounted.

2. The chip positioning method of the bulk chip placement machine according to claim 1, characterized in that: The correcting the motion path of the patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship includes: Calculating a deviation between the actual position relationship and the ideal position relationship, and obtaining a deviation between the position to be patched and the ideal patch position according to the deviation; Obtaining an X-axis deviation value and a Y-axis deviation value from the deviation between the position to be mounted and the ideal position of the patch; Based on the X-axis deviation value and the Y-axis deviation value, the motion path of the placement head of the bulk placement machine is corrected.

3. The chip positioning method of the bulk chip placement machine according to claim 1, characterized in that: The acquiring characteristic information of the circuit board and determining the corresponding image processing model based on the characteristic information of the circuit board includes: Acquire an overall image of the circuit board; Identifying circuit layout information and key device information from the overall image; Obtaining specific type information of the circuit board based on the circuit layout information and key component information; Based on the mapping relationship between the type information and the recognition model stored in the database, the recognition model corresponding to the specific type information is obtained as the image processing model.

4. A patch positioning device for a bulk patch machine, used to implement the method described in any one of claims 1 to 3 to achieve circuit board patch positioning, characterized in that: include: A first acquisition unit, which acquires characteristic information of a circuit board and determines a corresponding image processing model based on the characteristic information of the circuit board; An identification unit is used to start a camera on the bulk placement machine to collect an image of the circuit board, analyze the image based on the image processing model, and identify the actual position relationship between the reference mark point on the circuit board and the position to be placed; A second acquisition unit, used for acquiring an ideal position relationship between the reference mark point and the position to be mounted; A positioning unit, used for correcting the motion path of the patch head of the bulk patch machine according to the actual position relationship and the ideal position relationship; Based on the corrected motion path, the patch head is controlled to be positioned at the position to be patched.

5. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images