Circuit board defect detection method and system applying machine vision technology

By comprehensively evaluating multiple modules of the circuit board detection system, using machine vision technology and intelligent modules, the false alarm and missed detection problems of circuit board defect detection in complex environments are solved, and more efficient detection and alarm response are achieved, improving Quality management and efficiency of production lines.

CN119064378AInactive Publication Date: 2024-12-03HANGZHOU YUANSANG TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410549491.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-05-06
Publication Date
2024-12-03
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing circuit board defect detection system has false alarms or missed detection problems for complex and minor defects in complex industrial environments, especially on high-speed production lines, and has high computing resources and time requirements, which limits real-time.

Method used

A circuit board defect detection system using machine vision technology was designed. Through the comprehensive evaluation of assembly line performance detection module, camera performance detection module, environment detection module, data sorting module, data analysis module and identification module, transportation performance parameters and shooting are generated. Performance parameters and environmental parameters are used to obtain defect alarm parameters using dimensionless processing and integrated calculations, and compare the alarms through preset thresholds to reduce the false alarm rate.

Benefits of technology

Improve detection accuracy and alarm response speed, provide a more comprehensive and efficient circuit board defect detection solution, improve the quality management level and production efficiency of the production line, reduce false alarm rate, and improve the reliability and stability of the system. sex.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119064378A_ABST
    Figure CN119064378A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of machine vision, and discloses a circuit board defect detection method and system applying a machine vision technology, and the system comprises an assembly line performance detection module, a camera performance detection module, an environment detection module, a data arrangement module, a data analysis module, and an identification module. By integrating a plurality of modules such as an assembly line performance detection module, a camera performance detection module and an environment detection module, carrying out comprehensive evaluation by utilizing data arrangement and data analysis, and finally realizing defect alarm through an identification module, compared with a traditional circuit board detection method, the system realizes all-around performance monitoring, and the detection efficiency is improved. Compared with the prior art, the detection accuracy and the alarm response speed are improved, and compared with the current technical means, the system provides a more comprehensive and more efficient circuit board defect detection solution by utilizing the integration of a machine vision technology and an intelligent module, and the quality management level and the production efficiency of a production line are further improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of machine vision technology, specifically to a method and system for detecting defects in circuit boards using machine vision technology. Background Technology

[0002] Circuit board defect detection is a key area of ​​application of machine vision technology, belonging to the broad field of electronic manufacturing and quality control. In the modern electronics industry, highly integrated circuit boards play a vital role, and their quality directly affects the performance and reliability of electronic products. Machine vision technology is widely used in this field to improve production efficiency and ensure product quality, aiming to improve the quality control level of the production line by detecting various defects on circuit boards in an automated and intelligent manner.

[0003] Current visual inspection systems are often sensitive to factors such as lighting, angle, and scale, which may lead to unstable performance in complex industrial environments. For complex and tiny defects, existing systems may exhibit certain false alarms or missed detections, especially on high-speed production lines. In addition, large-scale data processing and model training may require considerable computing resources and time, which limits the ability to meet real-time requirements. Summary of the Invention

[0004] This invention provides a circuit board defect detection system and its usage method that utilizes machine vision technology, solving the problem mentioned in the background that existing systems may exhibit false alarms or missed detections for complex and minute defects, especially on high-speed production lines.

[0005] This invention provides the following technical solution: a circuit board defect detection system using machine vision technology, comprising a production line performance testing module, a camera performance testing module, an environmental monitoring module, a data processing module, a data analysis module, and an identification module.

[0006] The assembly line performance detection module is used to collect the operation performance data of the assembly line and input the collected operation performance data into the data processing module. The data processing module preprocesses the operation performance data and extracts features from the preprocessed operation performance data to generate the first data set.

[0007] The camera performance detection module is used to acquire the shooting effect data of the camera and input the acquired shooting effect data into the data processing module. The data processing module preprocesses the shooting effect data and extracts features from the preprocessed shooting effect data to generate a second data set.

[0008] The environmental detection module is used to collect transportation environment data of the circuit board and input the acquired transportation environment data into the data processing module. The data processing module preprocesses the transportation environment data and extracts features from the preprocessed transportation environment data to generate a third data set.

[0009] The data analysis module performs dimensionless processing on the first, second, and third data groups and then integrates and calculates them to generate transportation performance parameter YSxnc, shooting performance parameter PSxnc, and environmental parameter HJcs. It then performs dimensionless processing on the transportation performance parameter YSxnc, shooting performance parameter PSxnc, and environmental parameter HJcs again and integrates and calculates them to obtain the defect alarm parameter QXbjz. The calculated defect alarm parameter QXbjz is then input into the identification module.

[0010] The identification module is used to compare the calculated defect alarm parameter QXbjz with a preset first threshold Y to generate a first comparison result. Based on the first comparison result, it determines whether to issue an alarm for the currently inspected circuit board. If the determination result is that no alarm is needed, the circuit board is transported normally. If the determination result is that an alarm is needed, the defect alarm parameter QXbjz is first integrated with the first threshold Y to generate an alarm intervention value BJjrz. The generated alarm intervention value BJjrz is then compared with a preset second threshold R to generate a second comparison result. Based on the second comparison result, the corresponding alarm method and processing scheme are executed.

[0011] As an optional solution of the circuit board defect detection system using machine vision technology described in this invention, the data processing module includes a preprocessing unit and a feature extraction unit;

[0012] The preprocessing unit is used to preprocess the operation performance data, shooting effect data and transportation environment data, and the feature extraction unit is used to extract features from the operation performance data, shooting effect data and transportation environment data to generate a first data group, a second data group and a third data group.

[0013] As an optional solution of the circuit board defect detection system using machine vision technology described in this invention, the first data group includes the transport speed value YSsd and the cargo density value HWmdz;

[0014] The transport speed value YSsd represents the transport speed of the assembly line, and the cargo density value HWmdz represents the interval between cargo placements on the assembly line.

[0015] The second data group includes image sharpness TXqxd, image magnification TXfdz, normal color value JBysz, abnormal color value EWysz, trace width value HJkdz, and trace length value HJcdz;

[0016] Image clarity TXqxd represents the imaging pixels of the camera; image magnification TXfdz represents the shooting magnification of the camera; normal color value JBysz represents the color value that the circuit board should have; abnormal color value EWysz represents the abnormal color value of the circuit board; trace width value HJkdz represents the width of the solder trace on the circuit board; trace length value HJcdz represents the length of the solder trace on the circuit board.

[0017] The third data set includes the ambient fog value HJwqz and the ambient light value HJgzz;

[0018] The ambient fog value HJwqz represents the actual fog concentration in the environment during the transportation of the circuit board, and the ambient light value HJgzz represents the actual light intensity in the environment during the transportation of the circuit board.

[0019] As an optional solution of the circuit board defect detection system using machine vision technology described in this invention, the production line performance detection module includes a production line speed detection unit and a cargo density detection unit.

[0020] The assembly line speed detection unit is used to obtain the transport speed value YSsd, and the cargo density detection unit is used to obtain the cargo density value HWmdz.

[0021] As an optional solution of the circuit board defect detection system using machine vision technology described in this invention, the camera performance detection module includes a texture recognition unit, a color recognition unit, and a trace detection unit.

[0022] The texture recognition unit is used to acquire the image clarity TXqxd and the image magnification TXfdz, the color recognition unit is used to acquire the normal color value JBysz and the abnormal color value EWysz of the circuit board, and the trace detection unit is used to acquire the trace width value HJkdz and the trace length value HJcdz of the soldering trace on the circuit board.

[0023] As an optional solution of the circuit board defect detection system using machine vision technology described in this invention, the environmental detection module includes a fog detection unit and a light detection unit, wherein the fog detection unit is used to acquire the environmental fog value HJwqz, and the light detection unit is used to acquire the environmental light value HJgzz.

[0024] As an optional solution of the circuit board defect detection system using machine vision technology described in this invention, the identification module includes a first comparison unit and a second comparison unit, wherein the first comparison unit is used to generate a first comparison result and the second comparison unit is used to generate a second comparison result.

[0025] The first comparison result is:

[0026] When the defect alarm parameter QXbjz ≤ the first threshold Y, it means that the current circuit board does not need to generate an alarm.

[0027] When the defect alarm parameter QXbjz > the first threshold Y, it means that the current circuit board needs to issue an alarm.

[0028] The second comparison result is:

[0029] When the alarm intervention value BJjrz ≤ the second threshold R, it means that the current circuit board is in the first-level alarm signal and is marked with an orange signal.

[0030] When the alarm intervention value BJjrz > the second threshold R, it indicates that the current circuit board is in the second-level alarm signal and is marked with a red signal.

[0031] As an optional solution to the circuit board defect detection system using machine vision technology described in this invention, the identification module further includes a storage unit, a matching unit, and an execution unit;

[0032] The storage unit is used to store the processing plan, including:

[0033] When the circuit board is in the first-level alarm signal, notify the quality inspection personnel to remove the circuit board marked with orange and conduct manual quality inspection. If no obvious defects are found, the orange mark is removed and the circuit board is transported normally. If there are obvious defects, the mark is changed to red and the circuit board is sent to a professional instrument for testing. After the testing is completed, the corresponding repairs are carried out.

[0034] When the circuit board is in the level 2 alarm signal, notify the quality inspection personnel to remove the circuit board with the red mark and send it to a professional instrument for testing. After the testing is completed, carry out the corresponding repairs.

[0035] The matching unit is used to match the second comparison result with the processing scheme stored in the storage unit;

[0036] The execution unit is used to execute the processing plan, generate alarm logs, and save and archive the alarm logs.

[0037] As an optional solution of the circuit board defect detection system using machine vision technology described in this invention, the data analysis module includes a first calculation unit, a second calculation unit, and a third calculation unit.

[0038] The first calculation unit calculates and obtains the transportation performance parameter YSxnc, the shooting performance parameter PSxnc, and the environmental parameter HJcs using the following calculation formulas respectively;

[0039]

[0040]

[0041]

[0042] In the formula: YSsd is the transport speed value, HWmdz is the cargo density value, TXqxd is the image clarity, TXfdz is the image magnification, JBysz is the normal color value, EWysz is the abnormal color value, HJkdz is the trace width value, HJcd is the trace length value, HJwqz is the ambient fog value, and HJgzz is the ambient light value;

[0043] The second calculation unit calculates and obtains the defect alarm parameter QXbjz using the following calculation formula;

[0044]

[0045] YSxnc represents the transportation performance parameter, PSxnc represents the shooting performance parameter, HJcs represents the environmental parameter, a, a, and a are all equal, and a ≠ a ≠ a ≠, A is the first correction constant, and the values ​​of a, a, a, a, and A are adjusted and set by the user.

[0046] The third calculation unit calculates and obtains the alarm intervention value BJjrz using the following calculation formula.

[0047]

[0048] In the formula: QXbjz is the defect alarm value, Y is the first threshold, and B is the second correction constant, and the value of B is adjusted and set by the user.

[0049] This invention also provides a method for using the circuit board defect detection system using machine vision technology mentioned in the above specification, the specific steps of which are as follows;

[0050] S1. Collect data on the performance of the circuit board transport assembly line, the camera, and the circuit board transport environment to generate the first dataset, the second dataset, and the third dataset, respectively.

[0051] S2. Integrate and calculate the generated first dataset, second dataset, and third dataset to generate transportation performance parameter YSxnc, shooting performance parameter PSxnc, and environmental parameter HJcs. Integrate and calculate the transportation performance parameter YSxnc, shooting performance parameter PSxnc, and environmental parameter HJcs again to obtain the defect alarm parameter QXbjz.

[0052] S3. Compare the defect alarm parameter QXbjz with the preset first threshold Y to obtain the first comparison result. Integrate the defect alarm parameter QXbjz with the first threshold Y to obtain the alarm intervention value BJjrz. Compare the alarm intervention value BJjrz with the preset second threshold R to obtain the second comparison result.

[0053] S4. Based on the first comparison result, determine whether to issue an alarm for the circuit board being tested; based on the second comparison result, determine the specific alarm method and handling plan, and generate a processing log.

[0054] The present invention has the following beneficial effects:

[0055] 1. This circuit board defect detection system, which utilizes machine vision technology, integrates multiple modules such as production line performance testing, camera performance testing, and environmental monitoring. It uses data processing and analysis for comprehensive evaluation and finally achieves defect alarms through the identification module. Compared with traditional circuit board detection methods, this system achieves comprehensive performance monitoring, improves detection accuracy and alarm response speed. Compared with current technologies, this system, by integrating machine vision technology and intelligent modules, provides a more comprehensive and efficient circuit board defect detection solution, further improving the quality management level and production efficiency of the production line.

[0056] 2. This circuit board defect detection system, which utilizes machine vision technology, can more accurately distinguish alarm states under different conditions by setting two thresholds, namely the first threshold Y and the second threshold R. This reduces the false alarm rate, avoids unnecessary alarms caused by excessive system sensitivity, and improves the reliability and stability of the system. It also enables quality inspectors to make quick decisions and take corresponding measures, thereby improving the efficiency and accuracy of problem handling. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the system structure of the present invention.

[0058] Figure 2 This is a schematic diagram of the method flow of the present invention.

[0059] In the diagram: 101. Production line performance testing module; 102. Camera performance testing module; 103. Environmental detection module; 104. Data processing module; 105. Data analysis module; 106. Identification module; 201. Production line speed detection unit; 202. Cargo density detection unit; 204. Texture recognition unit; 205. Color recognition unit; 206. Trace detection unit; 207. Fog detection unit; 208. Illumination detection unit; 209. Preprocessing unit; 210. Feature extraction unit; 211. First calculation unit; 212. Second calculation unit; 213. Third calculation unit; 214. First comparison unit; 215. Second comparison unit; 216. Storage unit; 217. Matching unit; 218. Execution unit. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Example 1

[0062] Please see Figure 1 A circuit board defect detection system using machine vision technology is disclosed, including a production line performance testing module 101, a camera performance testing module 102, an environmental detection module 103, a data processing module 104, a data analysis module 105, and an identification module 106.

[0063] The production line performance detection module 101 is used to collect the operation performance data of the production line and input the collected operation performance data into the data processing module 104. The data processing module 104 preprocesses the operation performance data and extracts features from the preprocessed operation performance data to generate the first data set.

[0064] The camera performance detection module 102 is used to acquire the shooting effect data of the camera and input the acquired shooting effect data into the data processing module 104. The data processing module 104 preprocesses the shooting effect data and extracts features from the preprocessed shooting effect data to generate a second data set.

[0065] The environmental detection module 103 is used to collect transportation environment data of the circuit board and input the acquired transportation environment data into the data processing module 104. The data processing module 104 preprocesses the transportation environment data and extracts features from the preprocessed transportation environment data to generate a third data set.

[0066] The data analysis module 105 performs dimensionless processing on the first, second, and third data groups and then integrates and calculates them to generate the transportation performance parameter YSxnc, the shooting performance parameter PSxnc, and the environmental parameter HJcs. It then performs dimensionless processing on the transportation performance parameter YSxnc, the shooting performance parameter PSxnc, and the environmental parameter HJcs again and integrates and calculates them to obtain the defect alarm parameter QXbjz. Finally, it inputs the calculated defect alarm parameter QXbjz into the identification module 106.

[0067] The identification module 106 is used to compare the calculated defect alarm parameter QXbjz with a preset first threshold Y to generate a first comparison result. Based on the first comparison result, it determines whether to issue an alarm for the currently inspected circuit board. If the determination result is that no alarm is needed, the circuit board is transported normally. If the determination result is that an alarm is needed, the defect alarm parameter QXbjz is first integrated with the first threshold Y to generate an alarm intervention value BJjrz. The generated alarm intervention value BJjrz is then compared with a preset second threshold R to generate a second comparison result. Based on the second comparison result, the corresponding alarm method and processing scheme are executed.

[0068] In this embodiment: the production line performance detection module 101 is responsible for collecting production line operation performance data. It generates a first data set through preprocessing and feature extraction. By acquiring production line operation performance data, the accuracy of circuit board defect detection is improved.

[0069] The camera performance testing module 102 is used to collect the shooting effect data of the camera. It generates a second data set through preprocessing and feature extraction. Optimizing camera performance helps to improve image clarity, color accuracy, etc., thereby improving the accuracy of defect detection.

[0070] The environmental monitoring module 103 is responsible for collecting the transportation environment data of the circuit board. Through preprocessing and feature extraction, a third data set is generated to understand the environmental conditions faced by the circuit board during transportation. Optimizing environmental monitoring helps to improve the transportation safety of the circuit board and reduce the occurrence of damage and defects during transportation.

[0071] The data processing module 104 receives data from the pipeline performance testing module, camera performance testing module, and environmental monitoring module, preprocesses it, and extracts features to generate a comprehensive dataset. This helps integrate information from different modules, providing high-quality input for subsequent data analysis. Data processing improves data quality, reduces noise and redundant information, and thus enhances the accuracy of subsequent data analysis.

[0072] The data analysis module 105 performs dimensionless processing and comprehensive calculations on the integrated data to generate transportation performance parameters YSxnc, imaging performance parameters PSxnc, and environmental parameters HJcs. This enables a comprehensive evaluation of the circuit board's transportation, imaging, and environmental performance from multiple dimensions. Through data analysis, it can provide strong support for subsequent alarm decisions.

[0073] The identification module 106 is responsible for comparing the calculated defect alarm parameter QXbjz with the preset first threshold Y and executing the corresponding processing scheme, thereby realizing timely alarm and processing of defective circuit boards, improving the efficiency of defect detection. Through the intelligent identification module, timely response to alarms can be achieved, reducing the defect rate and improving the overall quality and efficiency of the production line.

[0074] This circuit board defect detection system integrates multiple modules such as production line performance testing, camera performance testing, and environmental monitoring. It utilizes data processing and analysis for comprehensive evaluation and ultimately uses an identification module to trigger a defect alarm. Compared to traditional circuit board inspection methods, this system achieves comprehensive performance monitoring, improves detection accuracy and alarm response speed. Compared to current technologies, this system, through the integration of machine vision technology and intelligent modules, provides a more comprehensive and efficient circuit board defect detection solution, further enhancing the quality management level and production efficiency of the production line.

[0075] Example 2

[0076] Please see Figure 1 The data processing module 104 includes a preprocessing unit 209 and a feature extraction unit 210;

[0077] The preprocessing unit 209 is used to preprocess the operation performance data, shooting effect data and transportation environment data, and the feature extraction unit 210 is used to extract features from the operation performance data, shooting effect data and transportation environment data to generate the first data group, the second data group and the third data group.

[0078] In this embodiment, the preprocessing unit 209 processes the operational performance data, shooting effect data, and transportation environment data, eliminating noise and redundant information in the data, thereby improving data quality. This makes subsequent data analysis more accurate and reliable, avoiding misjudgments and inaccurate conclusions caused by poor data quality.

[0079] The feature extraction unit 210 integrates multi-dimensional information on operational performance, imaging effects, and transportation environment by extracting features from data from different sources. This helps to understand the performance status of the circuit board from a more comprehensive perspective and improves the comprehensiveness and accuracy of defect detection.

[0080] The feature extraction unit 210 generates a first data set, a second data set, and a third data set, ensuring that each data set contains important features of the corresponding module. This customized data set can be applied more specifically to subsequent analysis and computation, thereby improving the efficiency of the entire system.

[0081] Through optimized preprocessing and feature extraction, the integrated data is more compact, reducing the system's data processing time. This helps improve the system's response speed, making defect detection and alarms react more quickly, thereby reducing the risks posed by potential production line problems.

[0082] Example 3

[0083] Please see Figure 1 The first data set includes the transport speed value YSsd and the cargo density value HWmdz;

[0084] The transport speed value YSsd represents the transport speed of the assembly line, and the cargo density value HWmdz represents the interval between cargo placements on the assembly line.

[0085] The second data set includes image sharpness TXqxd, image magnification TXfdz, normal color value JBysz, abnormal color value EWysz, trace width value HJkdz, and trace length value HJcdz.

[0086] Image clarity TXqxd represents the imaging pixels of the camera; image magnification TXfdz represents the shooting magnification of the camera; normal color value JBysz represents the color value that the circuit board should have; abnormal color value EWysz represents the abnormal color value of the circuit board; trace width value HJkdz represents the width of the solder trace on the circuit board; trace length value HJcdz represents the length of the solder trace on the circuit board.

[0087] The third data set includes the ambient fog value HJwqz and the ambient light value HJgzz;

[0088] The ambient fog value HJwqz represents the actual fog concentration in the environment during the transportation of the circuit board, and the ambient light value HJgzz represents the actual light intensity in the environment during the transportation of the circuit board.

[0089] The assembly line performance testing module 101 includes an assembly line speed testing unit 201 and a cargo density testing unit 202;

[0090] The assembly line speed detection unit 201 is used to obtain the transport speed value YSsd, and the cargo density detection unit 202 is used to obtain the cargo density value HWmdz.

[0091] The camera performance detection module 102 includes a texture recognition unit 204, a color recognition unit 205, and a trace detection unit 206;

[0092] The texture recognition unit 204 is used to obtain the image clarity TXqxd and the image magnification TXfdz, the color recognition unit 205 is used to obtain the normal color value JBysz and the abnormal color value EWysz of the circuit board, and the trace detection unit 206 is used to obtain the trace width value HJkdz and the trace length value HJcdz of the soldering trace on the circuit board.

[0093] The environmental detection module 103 includes a fog detection unit 207 and a light detection unit 208, wherein the fog detection unit 207 is used to acquire the environmental fog value HJwqz, and the light detection unit 208 is used to acquire the environmental light value HJgzz.

[0094] Example 4

[0095] Please see Figure 1 The identification module 106 includes a first comparison unit 214 and a second comparison unit 215. The first comparison unit 214 is used to generate a first comparison result, and the second comparison unit 215 is used to generate a second comparison result.

[0096] The first comparison result is:

[0097] When the defect alarm parameter QXbjz ≤ the first threshold Y, it means that the current circuit board does not need to generate an alarm.

[0098] When the defect alarm parameter QXbjz > the first threshold Y, it means that the current circuit board needs to issue an alarm.

[0099] The second comparison result is:

[0100] When the alarm intervention value BJjrz ≤ the second threshold R, it means that the current circuit board is in the first-level alarm signal and is marked with an orange signal.

[0101] When the alarm intervention value BJjrz > the second threshold R, it indicates that the current circuit board is in the second-level alarm signal and is marked with a red signal.

[0102] In this embodiment: the first comparison unit 214 enables the system to determine in real time whether the current circuit board needs to be alarmed. By comparing with the first threshold Y, the system can quickly decide whether to start the alarm process, thereby realizing timely response to potential defects and reducing the occurrence of abnormal events in the production process.

[0103] The second comparison unit 215 introduces multi-level alarm signals, comparing the alarm intervention value BJjrz with the second threshold R, and dividing them into first-level and second-level alarm signals. This multi-level alarm signal design allows the system to adopt different marking methods according to the urgency of the alarm, enabling maintenance personnel to more clearly understand the urgency of the current problem and helping to handle it more effectively.

[0104] By setting two thresholds, namely the first threshold Y and the second threshold R, the system can more accurately distinguish the alarm status under different conditions, thereby reducing the false alarm rate, avoiding unnecessary alarms caused by excessive system sensitivity, and thus improving the reliability and stability of the system. It also enables quality inspectors to make quick decisions and take corresponding measures, improving the efficiency and accuracy of problem handling.

[0105] Example 5

[0106] Please see Figure 1 The identification module 106 also includes a storage unit 216, a matching unit 217, and an execution unit 218;

[0107] Storage unit 216 is used to store processing plans, including:

[0108] When the circuit board is in the first-level alarm signal, notify the quality inspection personnel to remove the circuit board marked with orange and conduct manual quality inspection. If no obvious defects are found, the orange mark is removed and the circuit board is transported normally. If there are obvious defects, the mark is changed to red and the circuit board is sent to a professional instrument for testing. After the testing is completed, the corresponding repairs are carried out.

[0109] When the circuit board is in the level 2 alarm signal, notify the quality inspection personnel to remove the circuit board with the red mark and send it to a professional instrument for testing. After the testing is completed, carry out the corresponding repairs.

[0110] The matching unit 217 is used to match the second comparison result with the processing scheme stored in the storage unit 216;

[0111] The execution unit 218 is used to execute the processing plan, generate alarm logs, and save and archive the alarm logs.

[0112] In this embodiment, the storage unit 216 enables the system to save the processing schemes for alarms of different levels. This intelligent processing scheme management ensures that targeted processing measures can be taken in different situations, improving the applicability and flexibility of the system. The processing schemes in the storage unit 216 describe the quality inspection process in detail, including the specific processing methods for first-level alarms and second-level alarms. This helps to improve the standardization of operations for quality inspectors and ensures that corresponding actions can be taken quickly and accurately when an alarm occurs.

[0113] The matching unit 217 enables the system to intelligently match the second comparison result with the processing scheme in the storage unit 216. This matching mechanism ensures that the system can accurately select the appropriate processing scheme, reduces the probability of false alarms and false negatives, and improves the reliability of the system.

[0114] The execution unit 218 is responsible for executing the processing plan based on the matching results and generating corresponding alarm logs. This helps to achieve efficient alarm response, shorten the problem handling time, reduce the impact on the production line, and improve overall production efficiency.

[0115] The alarm logs generated by execution unit 218 are saved and archived, which facilitates post-event analysis and traceability. By recording the handling process of alarm events in detail, strong data support can be provided for quality management, helping the system to continuously improve and optimize.

[0116] Example 6

[0117] Please see Figure 1 The data analysis module 105 includes a first computing unit 211, a second computing unit 212, and a third computing unit 213;

[0118] The first calculation unit 211 calculates and obtains the transportation performance parameter YSxnc, the shooting performance parameter PSxnc, and the environmental parameter HJcs respectively using the following calculation formulas;

[0119]

[0120]

[0121]

[0122] In the formula: YSsd is the transport speed value, HWmdz is the cargo density value, TXqxd is the image clarity, TXfdz is the image magnification, JBysz is the normal color value, EWysz is the abnormal color value, HJkdz is the trace width value, HJcd is the trace length value, HJwqz is the ambient fog value, and HJgzz is the ambient light value;

[0123] The second calculation unit 212 calculates and obtains the defect alarm parameter QXbjz using the following calculation formula;

[0124]

[0125] YSxnc represents the transportation performance parameter, PSxnc represents the shooting performance parameter, HJcs represents the environmental parameter, a1, a2, and a3 are all constants, and a1≠a2≠a3≠0. A is the first correction constant, and the values ​​of a1, a2, a3, and A are adjusted and set by the user.

[0126] The third calculation unit 213 calculates and obtains the alarm intervention value BJjrz using the following calculation formula.

[0127]

[0128] In the formula: QXbjz is the defect alarm value, Y is the first threshold, and B is the second correction constant, and the value of B is adjusted and set by the user.

[0129] Example 7

[0130] This invention also provides a method for using the circuit board defect detection system employing machine vision technology mentioned in the above specification. Please refer to [link to relevant documentation]. Figure 2 The specific steps are as follows:

[0131] S1. Collect data on the performance of the circuit board transport assembly line, the camera, and the circuit board transport environment to generate the first dataset, the second dataset, and the third dataset, respectively.

[0132] S2. Integrate and calculate the generated first dataset, second dataset, and third dataset to generate transportation performance parameter YSxnc, shooting performance parameter PSxnc, and environmental parameter HJcs. Integrate and calculate the transportation performance parameter YSxnc, shooting performance parameter PSxnc, and environmental parameter HJcs again to obtain the defect alarm parameter QXbjz.

[0133] S3. Compare the defect alarm parameter QXbjz with the preset first threshold Y to obtain the first comparison result. Integrate the defect alarm parameter QXbjz with the first threshold Y to obtain the alarm intervention value BJjrz. Compare the alarm intervention value BJjrz with the preset second threshold R to obtain the second comparison result.

[0134] S4. Based on the first comparison result, determine whether to issue an alarm for the circuit board being tested. Based on the second comparison result, determine the specific alarm method and handling plan, and generate a processing log.

[0135] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0136] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A circuit board defect detection system using machine vision technology, comprising a production line performance detection module (101), a camera performance detection module (102), an environment detection module (103), a data sorting module (104), a data analysis module (105) and an identification module (106), characterized in that: The pipeline performance detection module (101) is used to collect the operation performance data of the pipeline, and input the collected operation performance data into the data sorting module (104), the data sorting module (104) pre-processes the operation performance data, and extracts features from the pre-processed operation performance data, thereby generating a first data group; The camera performance detection module (102) is used to acquire the camera's shooting effect data, and input the acquired shooting effect data into the data sorting module (104), pre-process the shooting effect data through the data sorting module (104), and perform feature extraction on the pre-processed shooting effect data, thereby generating a second data group; The environment detection module (103) is used to collect the transportation environment data of the circuit board, and input the acquired transportation environment data into the data sorting module (104), and the transportation environment data is preprocessed by the data sorting module (104), and feature extraction is performed on the preprocessed transportation environment data, thereby generating a third data group; The data analysis module (105) is used to perform dimensionless processing on the first data group, the second data group and the third data group and then perform integrated calculations to generate a transportation performance parameter YSxnc, a shooting performance parameter PSxnc and an environmental parameter HJcs, and perform dimensionless processing on the transportation performance parameter YSxnc, the shooting performance parameter PSxnc and the environmental parameter HJcs and then perform integrated calculations again to obtain a defect alarm parameter QXbjz, and input the calculated defect alarm parameter QXbjz into the identification module (106). The identification module (106) is used to compare the defect alarm parameter QXbjz obtained by calculation with a preset first threshold value Y, thereby generating a first comparison result, and judging whether to alarm the currently detected circuit board according to the first comparison result. If the judgment result is that no alarm is required, the circuit board is transported normally. If the judgment result is that an alarm is required, the defect alarm parameter QXbjz and the first threshold value Y are first integrated and calculated to generate an alarm intervention value BJjrz, and the generated alarm intervention value BJjrz is compared with a preset second threshold value R to generate a second comparison result, and according to the second comparison result, a corresponding alarm method and processing plan are executed.

2. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The data sorting module (104) includes a preprocessing unit (209) and a feature extraction unit (210); The preprocessing unit (209) is used to preprocess the operation performance data, the shooting effect data and the transportation environment data, and the feature extraction unit (210) is used to extract features from the operation performance data, the shooting effect data and the transportation environment data to generate a first data group, a second data group and a third data group.

3. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The first data set includes a transport speed value YSsd and a cargo density value HWmdz; The transport speed value YSsd represents the transport speed of the assembly line, and the cargo density value HWmdz represents the intervals at which cargo is placed on the assembly line; The second data group includes image clarity TXqxd, image magnification TXfdz, normal color value JBysz, abnormal color value EWysz, trace width value HJkdz and trace length value HJcdz; Image clarity TXqxd represents the imaging pixel of the camera, image magnification TXfdz represents the shooting magnification of the camera, normal color value JBysz represents the color value that the circuit board should have, abnormal color value EWysz represents the abnormal color value of the circuit board, trace width value HJkdz represents the width of the trace welded on the circuit board, and trace length value HJcdz represents the length of the trace welded on the circuit board; The third data set includes an ambient fog value HJwqz and an ambient light value HJgzz; The ambient fog value HJwqz represents the actual fog concentration of the environment in which the circuit board is transported, and the ambient light value HJgzz represents the actual light intensity of the environment in which the circuit board is transported.

4. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The assembly line performance detection module (101) comprises an assembly line speed detection unit (201) and a cargo density detection unit (202); The assembly line speed detection unit (201) is used to obtain a transport speed value YSsd, and the cargo density detection unit (202) is used to obtain a cargo density value HWmdz.

5. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The camera performance detection module (102) comprises a texture recognition unit (204), a color recognition unit (205) and a trace detection unit (206); The texture recognition unit (204) is used to obtain image clarity TXqxd and image magnification TXfdz, the color recognition unit (205) is used to obtain a normal color value JBysz and an abnormal color value EWysz of a circuit board, and the trace detection unit (206) is used to obtain a trace width value HJkdz and a trace length value HJcdz of welding on the circuit board.

6. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The environment detection module (103) comprises a fog detection unit (207) and a light detection unit (208), wherein the fog detection unit (207) is used to obtain an environment fog value HJwqz, and the light detection unit (208) is used to obtain an environment light value HJgzz.

7. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The identification module (106) comprises a first comparison unit (214) and a second comparison unit (215), wherein the first comparison unit (214) is used to generate a first comparison result, and the second comparison unit (215) is used to generate a second comparison result; The first comparison result is: When the defect alarm parameter QXbjz≤the first threshold value Y, it means that no alarm is required for the current circuit board; When the defect alarm parameter QXbjz> the first threshold value Y, it means that the current circuit board needs to alarm; The second comparison result is: When the alarm intervention value BJjrz≤the second threshold R, it means that the current circuit board is a first-level alarm signal and is marked with an orange signal; When the alarm intervention value BJjrz> the second threshold R, it means that the current circuit board is a level 2 alarm signal and is marked as a red signal.

8. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The identification module (106) further includes a storage unit (216), a matching unit (217) and an execution unit (218); The storage unit (216) is used to store the treatment plan, including: When the circuit board is a first-level alarm signal, notify the quality inspector to remove the circuit board with orange mark for manual quality inspection. If the inspection finds no obvious defect characteristics, the orange mark will be cancelled and the circuit board will be transported normally. If there are obvious defect characteristics, the red mark will be changed and sent to professional instruments for inspection. After the inspection is completed, the corresponding repair will be carried out; When the circuit board shows a level 2 warning signal, notify the quality inspector, remove the circuit board with a red mark, send it to a professional instrument for testing, and perform corresponding repairs after the test is completed; The matching unit (217) is used to match the second comparison result with the processing solution stored in the storage unit (216); The execution unit (218) is used to execute the processing plan, generate an alarm log, and save and archive the alarm log.

9. The circuit board defect detection system using machine vision technology according to claim 1, characterized in that: The data analysis module (105) comprises a first calculation unit (211), a second calculation unit (212) and a third calculation unit (213); The first calculation unit (211) calculates and obtains the transportation performance parameter YSxnc, the shooting performance parameter PSxnc and the environmental parameter HJcs respectively through the following calculation formulas; Where: YSsd is the transport speed value, HWmdz is the cargo density value, TXqxd is the image clarity, TXfdz is the image magnification, JBysz is the normal color value, EWysz is the abnormal color value, HJkdz is the trace width value, HJcd is the trace length value, HJwqz is the ambient fog value, and HJgzz is the ambient light value; The second calculation unit (212) calculates and obtains the defect alarm parameter QXbjz by the following calculation formula; YSxnc is the transport performance parameter, PSxnc is the shooting performance parameter, HJcs is the environmental parameter, a1, a2 and a3 are all, and a1≠a2≠a3≠0, A is the first correction constant, and the values ​​of a1, a2, a3 and A are adjusted by the user The third calculation unit (213) calculates and obtains the alarm intervention value BJjrz by the following calculation formula: Wherein: QXbjz is the defect alarm value, Y is the first threshold, B is the second correction constant, and the value of B is adjusted and set by the user.

10. A method for using a circuit board defect detection system using machine vision technology, characterized in that: A circuit board defect detection system using machine vision technology comprising any one of claims 1 to 9, wherein the specific steps are as follows: S1, collecting data on the performance of the circuit board transportation line, the camera, and the circuit board transportation environment, thereby generating a first data set, a second data set, and a third data set respectively; S2, integrating and calculating the generated first data set, second data set and third data set to generate a transportation performance parameter YSxnc, a shooting performance parameter PSxnc and an environmental parameter HJcs, and integrating and calculating the transportation performance parameter YSxnc, the shooting performance parameter PSxnc and the environmental parameter HJcs again to obtain a defect alarm parameter QXbjz; S3, comparing the defect alarm parameter QXbjz with a preset first threshold value Y to obtain a first comparison result, integrating and calculating the defect alarm parameter QXbjz and the first threshold value Y to obtain an alarm intervention value BJjrz, and comparing the alarm intervention value BJjrz with a preset second threshold value R to obtain a second comparison result; S4. Determine whether to alarm the circuit board being inspected according to the first comparison result, determine the specific alarm method and processing plan according to the second comparison result, and generate a processing log.