Multi-Scale Circuit Board Defect Detection Under Real-Time Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing circuit board defect detection methods face challenges due to the complexity of circuit board images, which have large sizes with small defective parts and dense layouts, leading to low detection efficiency and accuracy, and high computational burdens that hinder real-time requirements.
Innovation Solution
A method utilizing a multi-scale detection network and a codec to determine first and second defect regions, enhancing accuracy by comparing circuit board images with standard images, and employing graph convolutional networks to analyze graph structure features for improved detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional manual visual inspection is used for circuit board defect detection, then the detection process is simple to implement, but the detection efficiency and accuracy are low
Solution Approach 1:
The patent replaces manual visual inspection with an automated image processing system that uses neural networks to detect defects. The system captures images of circuit boards and automatically analyzes them to identify defect regions, eliminating the need for manual inspection while significantly improving detection efficiency and accuracy.
Solution Approach 2:
The patent creates a standard image representing a defect-free circuit board and uses it as a reference for comparison. By comparing the captured image of the circuit board with this standard image, the system can automatically identify defect regions, enabling automated detection without requiring complex multi-stage inspection systems.
2Measurement precision
If deep learning techniques with convolutional neural networks are applied to automatically extract features and locate defect regions, then detection accuracy is improved, but computational burden increases
Solution Approach 1:
The patent extracts and compares only the defect regions by using a standard image as a reference. Instead of analyzing the entire circuit board image through computationally intensive deep learning processes, the system identifies and focuses on specific defect locations by comparing the captured image against the standard image, thereby reducing overall computational burden while maintaining high accuracy.
3Reliability
If the circuit board image is processed to identify defect regions, then detection capability is enhanced, but the time required for processing increases
Solution Approach 1:
The patent prepares a standard image representing a defect-free circuit board in advance. This pre-prepared standard image serves as a reference that enables rapid comparison during actual defect detection, eliminating the need for time-consuming real-time analysis and allowing the system to quickly identify defect regions by simply comparing the captured image against the pre-existing standard.
Data Source
AI summary
The present disclosure relates to a method, a device, and a computer program product for detecting a circuit board defect. The method includes acquiring a circuit board image of the circuit board. The method further includes determining a first defect region according to the circuit board image, wherein the first defect region indicates a location of a defect in the circuit board. The method further includes determining a second defect region according to the circuit board image and a standard image for the circuit board, wherein the second defect region indicates a location of a defect in the circuit board, and the standard image indicates a circuit board without any defect. The method further includes determining a defect region of the circuit board according to the first defect region and the second defect region. Accordingly, the accuracy of detecting the defect region on the circuit board can be improved.