Circuit board production intelligent online analysis monitoring system and method thereof

Abstract

The invention provides an intelligent online analysis and monitoring system and method for circuit board production, and the system comprises an image collection module, a multispectral imaging module, a three-dimensional scanning measurement module, a copper thickness test module, a data collection module, an intelligent analysis and recognition module, and a self-adaptive algorithm optimization module. The system is used for carrying out real-time image acquisition on each key station in the circuit board production process by adopting a high-resolution industrial camera. Through high-definition image acquisition, multispectral imaging, three-dimensional scanning and measurement, comprehensive data acquisition in the circuit board production process is realized, the problem of limited data acquisition range in the prior art is solved, accurate detection and classification of circuit board defects are realized by applying an intelligent analysis algorithm, and the production efficiency is improved. The accuracy and efficiency of defect detection are improved, automatic adjustment and optimization of algorithm parameters are achieved through the self-adaptive algorithm optimization module, it is ensured that the algorithm can continuously adapt to changes of the production environment, and the stability and reliability of the system are improved.

Description

Technical Field

[0001] The present invention relates to a monitoring system and method thereof, specifically an intelligent online analysis and monitoring system and method for circuit board production, belonging to the technical field of circuit board production. Background Art

[0002] The intelligent online analysis and monitoring system for circuit board production is a system integrating advanced technologies and algorithms, aiming to monitor and analyze various defects in the circuit board production process in real time. This system can efficiently and accurately identify defects such as open circuits, short circuits, line gaps, protrusions, and residual copper in products, thereby ensuring product quality and production efficiency. The intelligent online analysis and monitoring system for circuit board production generally includes the following main parts: an image acquisition module, an image processing and analysis module, a defect identification and classification module, and a data management and report generation module. These modules work together to achieve comprehensive monitoring of the circuit board production process. The system can monitor the circuit board production process in real time, promptly discover and repair defects, reduce the generation of defective products, and moreover, by reducing the costs of manual inspection and repair, the system can significantly reduce production costs. By supporting the detection of various types of circuit boards, the system can adapt to the production requirements of different products.

[0003] It is known that the Chinese published invention (Publication No.: CN105372581A) discloses an automatic monitoring and intelligent analysis system and method for the manufacturing process of flexible circuit boards. It uses automated devices and systems to conduct automated monitoring and analysis of the manufacturing process of flexible circuit boards. Compared with traditional manual inspection, it not only reduces the false alarm rate of inspection, increases the inspection categories, but also can greatly improve the production efficiency and automation level of flexible circuit boards. By introducing microscopes and precision electric platforms into the monitoring of key processes of flexible circuit boards, on the one hand, it effectively improves the accuracy of the system's monitoring objects, and on the other hand, it simplifies the inspection process and improves the inspection efficiency of the system. By using methods such as statistical process control and neural networks, it realizes intelligent quality monitoring and fault diagnosis, thereby more effectively ensuring the quality of flexible circuit board production and improving the ability of fault diagnosis during the production process;

[0004] Although it realizes the monitoring and intelligent analysis of key processes of circuit boards, it still has the following deficiencies:

[0005] Data acquisition limitations: It mainly relies on automatic data acquisition by microscopes and copper thickness testing devices. These two devices often cannot cover all key parameters and defects in the circuit board manufacturing process, and data acquisition is limited by the accuracy and resolution of the devices, resulting in the inability to accurately capture some minor or complex defects;

[0006] Intelligent analysis accuracy: The intelligent analysis algorithms of its system often cannot adapt to all types of defects and parameter changes, resulting in false alarms or missed alarms;

[0007] Insufficient algorithm adaptability: It lacks the adaptive ability for different types of circuit boards or different production conditions, resulting in poor analysis effects in certain specific situations. Moreover, the algorithm overly relies on historical data and ignores the dynamic changes of real-time data, leading to inaccurate analysis results.

[0008] Therefore, an intelligent on-line analysis and monitoring system and method for circuit board production are proposed. Summary of the Invention

[0009] In view of this, the present invention provides an intelligent on-line analysis and monitoring system and method for circuit board production to solve or alleviate the technical problems existing in the prior art and at least provide a beneficial option.

[0010] The technical solution of the embodiment of the present invention is implemented as follows: An intelligent on-line analysis and monitoring system for circuit board production includes an image acquisition module, a multi-spectral imaging module, a three-dimensional scanning and measurement module, a copper thickness test module, a data acquisition module, an intelligent analysis and recognition module, and an adaptive algorithm optimization module:

[0011] The image acquisition module is used to perform real-time image acquisition on each key station in the circuit board production process by using a high-resolution industrial camera. The key stations include an etching station, a drilling station, an electroplating station, a solder resist station, and an assembly station;

[0012] The multi-spectral imaging module is used to image the circuit board with light of different wavelengths to achieve precise identification of the surface material, structure, and defects of the circuit board;

[0013] The three-dimensional scanning and measurement module is used to perform three-dimensional measurement on the circuit board to obtain three-dimensional data on the surface of the circuit board and detect three-dimensional defects;

[0014] The copper thickness test module is used to measure the thickness of the copper layer of the circuit board by using a non-contact measurement method;

[0015] The data acquisition module is used to collect key information and status information in the production process. The key information includes process parameters, material information, and equipment status. The status information includes production progress information, quality monitoring information, and environmental parameters;

[0016] The intelligent analysis and recognition module is used to analyze and identify the collected images and data by using an intelligent analysis algorithm to detect and classify the defects of the circuit board;

[0017] The adaptive algorithm optimization module is used to automatically adjust and optimize the parameters of the intelligent analysis algorithm according to the data changes in the production process by using an adaptive optimization algorithm. Through real-time learning and updating the algorithm model, the intelligent analysis algorithm continuously adapts to the changes in the production environment.

[0018] Further preferably, the image acquisition module captures the circuit board images before and after etching at the etching station;

[0019] The image acquisition module records the position, size and depth of the drilled holes at the drilling station;

[0020] The image acquisition module monitors the uniformity and thickness of the electroplated layer at the electroplating station;

[0021] The image acquisition module checks the integrity and uniformity of the solder mask layer, as well as the cases of missing coating and uneven coating at the solder mask station;

[0022] The image acquisition module monitors the mounting position, orientation and integrity of the components at the assembly station.

[0023] Further preferably, the multispectral imaging module captures the color, texture and shape information of the circuit board surface through visible light, captures the fluorescent substances and defect information on the circuit board surface through ultraviolet light, captures the temperature distribution information on the circuit board surface through infrared light, and captures the internal structure and defect information of the circuit board through X-rays.

[0024] Further preferably, the non-contact measurement method includes:

[0025] Laser ranging: Using a laser beam to irradiate the surface of the copper layer of the circuit board, and calculating the thickness of the copper layer by measuring the reflection time or phase difference of the laser beam;

[0026] X-ray fluorescence spectrometry: Using X-rays to irradiate the copper layer of the circuit board, exciting the inner-layer electrons of copper atoms to transition and releasing characteristic fluorescent X-rays, and calculating the thickness of the copper layer by measuring the energy and intensity of the fluorescent X-rays.

[0027] Further preferably, the intelligent analysis algorithm includes the following steps:

[0028] Data preprocessing, performing enhancement processing on the collected circuit board images, including adjusting brightness, contrast, applying image filtering technology to remove noise, converting the image data into a unified size and format, and making corresponding defect labels for each image, including defect type, position, size, for algorithm training and verification;

[0029] Feature extraction, extracting high-level features from the circuit board images, and fusing features at different levels through the skip connection method;

[0030] Defect Detection and Classification: Based on feature extraction, a region proposal network is designed to generate defect regions, and the generated defect regions are classified and regressed;

[0031] Algorithm Optimization and Design of Loss Function: A loss function is designed to measure the gap between the algorithm prediction result and the true label. The loss function consists of two parts: classification loss and regression loss;

[0032] Model Training and Optimization: An intelligent analysis module is constructed. The model is trained using the labeled circuit board image data, and the network parameters are updated through the backpropagation algorithm to minimize the loss function;

[0033] Algorithm Evaluation and Verification: The labeled circuit board image data is divided into a training set, a validation set, and a test set for model training, validation, and testing.

[0034] Further preferably, the adaptive optimization algorithm includes the following steps:

[0035] Initialization: Select a convolutional neural network to construct an adaptive optimization model, initialize the model's weight and bias parameters, set the learning rate to control the speed of parameter update, and set the momentum factor to accelerate the optimization process;

[0036] Data Collection: Key data during the circuit board production process is obtained in real-time, and circuit board image data is collected simultaneously. The collected data is cleaned to remove noise and outliers, and the image data is normalized so that the pixel values of the processed images are within a unified range;

[0037] Algorithm Model Operation: The preprocessed data is input into the initialized algorithm model for feature extraction and classification. For image data, the model will output the prediction result of whether there are defects on the circuit board, as well as the type and location of the defects;

[0038] Calculation of Algorithm Performance Metrics: According to the model's prediction result and the true label, the performance metrics of the model are calculated;

[0039] Adaptive Adjustment: According to the performance metrics and the adaptive mechanism, the adjustment amount of the model parameters is calculated based on the calculated performance metrics and the preset adaptive mechanism. The magnitude and direction of the adjustment amount depend on the settings of the learning rate and the momentum factor. According to the calculated parameter adjustment amount, the parameter set of the algorithm model is updated, and the updated parameter set is used for data processing and prediction in the next iteration.

[0040] Further preferably, it further includes a defect classification and marking module, which is used to classify and mark the identified defects, including open circuits, short circuits, line gaps, and copper layer defects, and provides detailed defect information and location markings.

[0041] Further preferably, it further includes an early warning module, which is used to send an alarm signal and provide corresponding processing suggestions when a defect is detected. The processing suggestions include stopping the machine for inspection, adjusting process parameters, replacing raw materials, and manual re-inspection.

[0042] Further preferably, it further includes a data storage module, which is used to uniformly manage and store the collected data, including raw data, analysis results, and alarm records.

[0043] An intelligent online analysis and monitoring method for circuit board production includes the following steps:

[0044] Step 1: Use a high-resolution industrial camera to perform real-time image acquisition on each key station during the circuit board production process;

[0045] Step 2: Use light of different wavelengths to image the circuit board, and identify the surface material, structure, and defects of the circuit board;

[0046] Step 3: Perform three-dimensional measurement on the circuit board to obtain three-dimensional data on the surface of the circuit board, and detect three-dimensional defects;

[0047] Step 4: Use a non-contact measurement method to measure the thickness of the copper layer of the circuit board;

[0048] Step 5: Collect key information and status information during the production process, analyze and identify the collected images and data through intelligent analysis algorithms, and detect and classify the defects of the circuit board;

[0049] Step 6: Use an adaptive optimization algorithm to automatically adjust and optimize the parameters of the intelligent analysis algorithm according to the data changes during the production process. Through real-time learning and updating the algorithm model, the intelligent analysis algorithm continuously adapts to the changes in the production environment;

[0050] Step 7: Classify and mark the identified defects, including open circuits, short circuits, line gaps, and copper layer defects, and provide detailed defect information and position marks.

[0051] Due to the above technical solutions adopted in the embodiments of the present invention, it has the following advantages:

[0052] Through technologies such as high-definition image acquisition, multispectral imaging, three-dimensional scanning and measurement, the present invention realizes comprehensive data acquisition in the circuit board production process, solves the problem of limited data acquisition range in the prior art, realizes precise detection and classification of circuit board defects by using intelligent analysis algorithms, improves the accuracy and efficiency of defect detection, realizes automatic adjustment and optimization of algorithm parameters through an adaptive algorithm optimization module, ensures that the algorithm can continuously adapt to changes in the production environment, improves the stability and reliability of the system, and moreover, through a real-time alarm and feedback mechanism, timely reminds the operator to take measures, avoids defective products from flowing into the next process, ensures the continuity and stability of the production process. Compared with the prior art, the present invention solves the problems of large limitations in data acquisition, poor analysis accuracy, and insufficient algorithm adaptability.

[0053] The above summary is only for the purpose of the specification and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present invention will be readily apparent by reference to the drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0055] Figure 1 It is a schematic structural diagram of an intelligent on-line analysis and monitoring system for circuit board production according to the present invention;

[0056] Figure 2 It is a step flow chart of the intelligent analysis algorithm according to the present invention;

[0057] Figure 3 It is a step flow chart of the adaptive optimization algorithm according to the present invention;

[0058] Figure 4 It is a step flow chart of an intelligent on-line analysis and monitoring method for circuit board production according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] In the following, only some exemplary embodiments are simply described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the present invention. Therefore, the drawings and the description are considered to be exemplary in nature rather than restrictive.

[0060] The embodiments of the present invention will be described in detail below with reference to the drawings.

[0061] As Figures 1 - 4 shown, an intelligent on-line analysis and monitoring system for circuit board production provided by an embodiment of the present invention includes an image acquisition module, a multispectral imaging module, a three-dimensional scanning and measurement module, a copper thickness test module, a data acquisition module, an intelligent analysis and recognition module, and an adaptive algorithm optimization module;

[0062] The image acquisition module is used to perform real-time image acquisition on each key work station in the circuit board production process by using a high-resolution industrial camera. The key work stations include an etching work station, a drilling work station, an electroplating work station, a solder mask work station, and an assembly work station;

[0063] The image acquisition module captures images of the circuit board before and after etching at the etching work station;

[0064] The image acquisition module records the position, size, and depth of the drilled holes at the drilling work station;

[0065] The image acquisition module monitors the uniformity and thickness of the electroplated layer at the electroplating work station;

[0066] The image acquisition module checks the integrity and uniformity of the solder mask layer, as well as the cases of missing coating and uneven coating, at the solder mask work station;

[0067] The image acquisition module monitors the mounting position, direction, and integrity of components at the assembly work station.

[0068] The multispectral imaging module is used to image the circuit board using light of different wavelengths to achieve precise identification of the surface material, structure, and defects of the circuit board. The multispectral imaging module captures color, texture, and shape information on the surface of the circuit board through visible light, captures fluorescence substances and defect information on the surface of the circuit board through ultraviolet light, captures temperature distribution information on the surface of the circuit board through infrared light, and captures internal structure and defect information of the circuit board through X-rays;

[0069] In the circuit board detection, the multispectral imaging module realizes its function through the following steps:

[0070] Light source selection: First, it is necessary to select appropriate light sources according to the material, structure, and defect type of the circuit board. These light sources can usually emit light of different wavelengths, such as visible light, ultraviolet light, infrared light, etc.;

[0071] Imaging system: Next, use a high-precision camera and lens system to capture images of the circuit board under light of different wavelengths. These images will contain information such as reflection, transmission, and scattering on the surface of the circuit board;

[0072] Image processing: Perform preprocessing on the captured images, such as denoising and enhancing contrast, to improve the image quality. Then, use image processing algorithms to extract feature information on the surface of the circuit board, such as color, texture, shape, etc.;

[0073] Defect detection: Compare the extracted feature information with a preset defect library to identify defects on the circuit board. At the same time, classify and label the defects according to the type and severity of the defects;

[0074] Data analysis and feedback: Finally, perform data analysis on the detected defects to generate a detection report and provide it to the production personnel for processing. At the same time, the production process can be optimized and improved according to the detection results.

[0075] The 3D scanning measurement module is used to perform 3D measurement on the circuit board, obtain 3D data on the surface of the circuit board, and detect 3D defects;

[0076] The copper thickness testing module is used to measure the thickness of the copper layer on the circuit board by using a non-contact measurement method. The non-contact measurement methods include:

[0077] Laser ranging: Use a laser beam to irradiate the surface of the copper layer on the circuit board, and calculate the thickness of the copper layer by measuring the reflection time or phase difference of the laser beam;

[0078] X-ray fluorescence spectrometry: Use X-rays to irradiate the copper layer on the circuit board, excite the inner layer electrons of copper atoms to transition and release characteristic fluorescent X-rays, and calculate the thickness of the copper layer by measuring the energy and intensity of the fluorescent X-rays;

[0079] The copper thickness testing module usually consists of the following parts:

[0080] Light source / transmitter: Used to emit laser beams or X-rays;

[0081] Detector / receiver: Used to receive the reflected laser beam or fluorescent X-rays and convert them into electrical signals for processing;

[0082] Signal processing unit: Amplify, filter, and digitize the received electrical signals to extract information on the thickness of the copper layer;

[0083] Control unit: Responsible for controlling the operation of the entire measurement system, including the switching of the light source, the calibration of the detector, and data acquisition, etc.;

[0084] Display and storage unit: Used to display the measurement results and store the measurement data for subsequent analysis and processing;

[0085] The data acquisition module is used to collect key information and status information during the production process. The key information includes process parameters, material information, and equipment status. The status information includes production progress information, quality monitoring information, and environmental parameters;

[0086] The intelligent analysis and recognition module is used to analyze and recognize the collected images and data by using intelligent analysis algorithms, and detect and classify the defects of the circuit board;

[0087] The intelligent analysis algorithms include the following steps:

[0088] Data preprocessing: Enhance the collected circuit board images, including adjusting brightness and contrast, applying image filtering techniques to remove noise, converting the image data into a unified size and format, and creating corresponding defect labels for each image, including defect type, location, and size, for algorithm training and verification;

[0089] Feature extraction: Extract high-level features from the circuit board images and fuse features at different levels through skip connection methods;

[0090] Defect detection and classification: Based on feature extraction, design a region proposal network to generate defect regions and classify and regress the generated defect regions;

[0091] Algorithm optimization: Design a loss function to measure the gap between the algorithm prediction results and the true labels. The loss function consists of two parts: classification loss and regression loss;

[0092] Model training and optimization: Build an intelligent analysis module, train the model using the labeled circuit board image data, update the network parameters through the backpropagation algorithm, and minimize the loss function;

[0093] Algorithm evaluation and verification: Divide the labeled circuit board image data into a training set, a validation set, and a test set for model training, validation, and testing;

[0094] The adaptive algorithm optimization module is used to automatically adjust and optimize the parameters of the intelligent analysis algorithm according to the data changes in the production process. Through real-time learning and updating the algorithm model, the intelligent analysis algorithm continuously adapts to the changes in the production environment;

[0095] The adaptive optimization algorithm includes the following steps:

[0096] Initialization: Select a convolutional neural network to build an adaptive optimization model, initialize the model's weight and bias parameters, set the learning rate to control the speed of parameter update, and set the momentum factor to accelerate the optimization process;

[0097] Data collection: Real-time obtain the key data in the circuit board production process, and at the same time collect the circuit board image data. Clean the collected data to remove noise and outliers, and normalize the image data so that the pixel values of the processed images are within a unified range;

[0098] Algorithm model operation: Input the pre - processed data into the initialized algorithm model for feature extraction and classification. For image data, the model will output the prediction results of whether there are defects on the circuit board, as well as the types and locations of the defects.

[0099] Calculate the performance metrics of the algorithm: Calculate the performance metrics of the model based on the prediction results of the model and the true labels.

[0100] Adaptive adjustment: According to the performance metrics and the adaptive mechanism, calculate the adjustment amount of the model parameters based on the calculated performance metrics and the preset adaptive mechanism. The magnitude and direction of the adjustment amount depend on the settings of the learning rate and momentum factor. Update the parameter set of the algorithm model according to the calculated parameter adjustment amount. The updated parameter set is used for data processing and prediction in the next iteration.

[0101] Stability check: After updating the parameters, use the validation dataset to test the model and evaluate its stability. If the performance of the model shows a significant decline or fluctuation, a stability check is required. If it is unstable, roll back to the previous parameter set and adjust the parameters of the adaptive mechanism. If the model is unstable, roll back to the parameter set before the update and adjust the parameters of the adaptive mechanism, such as reducing the learning rate or increasing the momentum factor. The adjusted parameters will be used in the next iteration of the adaptive adjustment process.

[0102] Iteration and update: Repeat the above steps until the algorithm performance reaches the preset target or the number of iterations reaches the upper limit. During the iteration process, continuously collect new data, update the model parameters, and evaluate the performance of the model. When the performance of the model reaches the preset target or the number of iterations reaches the upper limit, stop the iteration process. During the iteration process, gradually optimize the parameters of the adaptive mechanism to improve the adaptive ability of the algorithm. During the iteration process, adjust the parameters of the adaptive mechanism, such as the learning rate and momentum factor, according to the performance changes of the model.

[0103] In one embodiment, it further includes a defect classification and marking module, which is used to classify and mark the identified defects, including open circuits, short circuits, line gaps, and copper layer defects, and provide detailed defect information and location markings.

[0104] In one embodiment, it further includes an early warning module. The early warning module is used to send an alarm signal and provide corresponding processing suggestions when a defect is detected. The processing suggestions include stopping the machine for inspection, adjusting process parameters, replacing raw materials, and manual re - inspection. Specifically:

[0105] Immediately stop the machine for inspection: If the defect is severe enough to potentially affect product quality or production safety, the system will recommend immediately stopping the machine to inspect the equipment or production line to prevent the production of more defective products.

[0106] Adjust process parameters: For defects that can be corrected by adjusting process parameters, the system will provide specific parameter adjustment suggestions, such as temperature, pressure, and speed;

[0107] Replace raw materials or components: If the defect is caused by the quality problem of raw materials or components, the system will recommend replacing qualified raw materials or components;

[0108] Manual re-inspection and repair: For some defects that can be solved by manual re-inspection and repair, the system will recommend sending the defective products to the designated area for manual processing.

[0109] In one embodiment, it further includes a data storage module, which is used to uniformly manage and store the collected data, including raw data, analysis results, and alarm records.

[0110] An intelligent online analysis and monitoring method for circuit board production includes the following steps:

[0111] Step 1: Use a high-resolution industrial camera to perform real-time image acquisition on each key station during the circuit board production process;

[0112] Step 2: Use light of different wavelengths to image the circuit board, and identify the surface material, structure, and defects of the circuit board;

[0113] Step 3: Perform three-dimensional measurement on the circuit board, obtain the three-dimensional data of the circuit board surface, and detect three-dimensional defects;

[0114] Step 4: Use a non-contact measurement method to measure the thickness of the copper layer of the circuit board;

[0115] Step 5: Collect key information and status information during the production process, analyze and identify the collected images and data through intelligent analysis algorithms, and detect and classify the defects of the circuit board;

[0116] Step 6: Use an adaptive optimization algorithm to automatically adjust and optimize the parameters of the intelligent analysis algorithm according to the data changes during the production process. Through real-time learning and updating the algorithm model, the intelligent analysis algorithm continuously adapts to the changes in the production environment;

[0117] Step 7: Classify and mark the identified defects, including open circuit, short circuit, line gap, and copper layer defect, and provide detailed defect information and location marking.

[0118] When the present invention is in operation: The image acquisition module performs real-time image acquisition on each key workstation during the circuit board production process. The multispectral imaging module uses light of different wavelengths to image the circuit board, identify the surface material, structure, and defects of the circuit board. The three-dimensional scanning and measurement module performs three-dimensional measurement on the circuit board to obtain the three-dimensional data of the circuit board surface and detect three-dimensional defects. The copper thickness test module measures the thickness of the copper layer of the circuit board. Then, the data acquisition module collects the key information and status information during the production process. The intelligent analysis algorithm analyzes and identifies the acquired images and data to detect and classify the defects of the circuit board. Then, the adaptive optimization algorithm automatically adjusts and optimizes the parameters of the intelligent analysis algorithm according to the data changes during the production process. The defect classification and marking module classifies and marks the identified defects, including open circuits, short circuits, line gaps, and copper layer defects, and provides detailed defect information and position markings.

[0119] The above is only the specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. An intelligent online analysis and monitoring system for circuit board production, comprising an image acquisition module, a multispectral imaging module, a three-dimensional scanning measurement module, a copper thickness test module, a data acquisition module, an intelligent analysis and recognition module, and an adaptive algorithm optimization module, characterized by: The image acquisition module is used to use a high-resolution industrial camera to perform real-time image acquisition of each key station in the circuit board production process, wherein the key stations include etching stations, drilling stations, electroplating stations, solder mask stations and assembly stations; The multispectral imaging module is used to image the circuit board using light of different wavelengths to achieve accurate identification of the material, structure, and defects on the circuit board surface; The three-dimensional scanning and measuring module is used to perform three-dimensional measurement of the circuit board, obtain three-dimensional data of the circuit board surface, and detect three-dimensional defects; The copper thickness testing module is used to measure the thickness of the copper layer of the circuit board using a non-contact measurement method; The data acquisition module is used to collect key information and status information during the production process. The key information includes process parameters, material information and equipment status. The status information includes production progress information, quality monitoring information and environmental parameters. The intelligent analysis and recognition module is used to analyze and recognize the collected images and data using an intelligent analysis algorithm, and detect and classify defects on the circuit board; The adaptive algorithm optimization module is used to automatically adjust and optimize the parameters of the intelligent analysis algorithm according to data changes in the production process using an adaptive optimization algorithm. By learning and updating the algorithm model in real time, the intelligent analysis algorithm continuously adapts to changes in the production environment.

2. The intelligent online analysis and monitoring system for circuit board production according to claim 1, wherein: The image acquisition module captures images of the circuit board before and after etching at the etching station; The image acquisition module records the position, size and depth of the drill hole at the drilling station; The image acquisition module monitors the uniformity and thickness of the electroplating layer at the electroplating station; The image acquisition module checks the integrity and uniformity of the solder mask layer at the solder mask station, as well as any missing or uneven coating conditions; The image acquisition module monitors the placement position, direction and integrity of components at the assembly station.

3. An intelligent on-line analysis and monitoring system for circuit board production according to claim 1, characterized in that: The multispectral imaging module captures the color, texture and shape information of the circuit board surface through visible light, captures the fluorescent substances and defect information on the circuit board surface through ultraviolet light, captures the temperature distribution information on the circuit board surface through infrared light, and captures the structure and defect information inside the circuit board through X-rays.

4. An intelligent online analysis and monitoring system for circuit board production according to claim 1, characterized in that: The non-contact measurement method comprises: Laser ranging: Use a laser beam to illuminate the copper surface of the circuit board, and calculate the thickness of the copper layer by measuring the reflection time or phase difference of the laser beam; X-ray fluorescence spectroscopy: X-rays are used to irradiate the copper layer of the circuit board, stimulating the inner electron transition of the copper atoms and releasing characteristic fluorescent X-rays. The thickness of the copper layer is calculated by measuring the energy and intensity of the fluorescent X-rays.

5. An intelligent online analysis and monitoring system for circuit board production according to claim 1, characterized in that: The intelligent analysis algorithm includes the following steps: Data preprocessing: Perform enhancement processing on the collected circuit board images, including adjusting brightness and contrast, applying image filtering techniques to remove noise, converting the image data into a unified size and format, and creating corresponding defect labels for each image, including defect type, location, and size, for the training and validation of the algorithm; Feature extraction: Extract high-level features from the circuit board images and fuse features at different levels through the skip connection method; Defect detection and classification: Based on feature extraction, design a region proposal network for generating defect regions and classify and regress the generated defect regions; Algorithm optimization: Design a loss function to measure the gap between the algorithm prediction result and the true label. The loss function consists of two parts: classification loss and regression loss; Model training and optimization: Construct an intelligent analysis module, use the labeled circuit board image data to train the model, update the network parameters through the backpropagation algorithm, and minimize the loss function; Algorithm evaluation and verification: Divide the labeled circuit board image data into a training set, a validation set, and a test set for model training, validation, and testing.

6. The intelligent on-line analysis and monitoring system for circuit board production according to claim 1, wherein: The adaptive optimization algorithm includes the following steps: Initialization: Select a convolutional neural network to construct an adaptive optimization model, initialize the weight and bias parameters of the model, set the learning rate to control the speed of parameter update, and set the momentum factor to accelerate the optimization process; Data collection: Real-time obtain key data during the circuit board production process, and at the same time collect circuit board image data. Clean the collected data to remove noise and outliers, and normalize the image data so that the pixel values of the processed images are within a unified range; Algorithm model operation: Input the preprocessed data into the initialized algorithm model for feature extraction and classification. For image data, the model will output the prediction result of whether there are defects on the circuit board, as well as the type and location of the defects; Calculate the performance metrics of the algorithm: Calculate the performance metrics of the model based on the prediction results of the model and the true labels; Adaptive adjustment: According to the performance metrics and the adaptive mechanism, calculate the adjustment amount of the model parameters based on the calculated performance metrics and the preset adaptive mechanism. The size and direction of the adjustment amount depend on the settings of the learning rate and the momentum factor. Update the parameter set of the algorithm model according to the calculated parameter adjustment amount, and the updated parameter set is used for data processing and prediction in the next iteration.

7. An intelligent online analysis and monitoring system for circuit board production according to claim 1, characterized in that: It also includes a defect classification and marking module, which is used to classify and mark the identified defects, including open circuits, short circuits, line gaps, and copper layer defects, and provide detailed defect information and location markings.

8. An intelligent online analysis and monitoring system for circuit board production according to claim 1, characterized in that: It also includes an early warning module, which is used to send an alarm signal when a defect is detected and provide corresponding processing suggestions. The processing suggestions include stopping for inspection, adjusting process parameters, replacing raw materials, and manual re-inspection.

9. An intelligent online analysis and monitoring system for circuit board production according to claim 1, characterized in that: It also includes a data storage module, which is used to uniformly manage and store the collected data, including raw data, analysis results, and alarm records.

10. A smart online analysis and monitoring method for circuit board production, applied to the smart online analysis and monitoring system for circuit board production according to any one of claims 1-9, characterized in that, Include the following steps: Step 1: Use a high-resolution industrial camera to perform real-time image acquisition at each key station during the circuit board production process; Step 2: Image the circuit board using light of different wavelengths to identify the surface material, structure, and defects of the circuit board; Step 3: Perform three-dimensional measurement on the circuit board to obtain three-dimensional data of the circuit board surface and detect three-dimensional defects; Step 4: Measure the thickness of the copper layer of the circuit board using a non-contact measurement method; Step 5: Collect key information and status information during the production process, analyze and identify the collected images and data through intelligent analysis algorithms, and detect and classify the defects of the circuit board; Step 6: Use an adaptive optimization algorithm to automatically adjust and optimize the parameters of the intelligent analysis algorithm according to the data changes during the production process. Through real-time learning and updating the algorithm model, the intelligent analysis algorithm continuously adapts to the changes in the production environment; Step 7: Classify and mark the identified defects, including open circuits, short circuits, line gaps, and copper layer defects, and provide detailed defect information and location markings.

Images