Electronic Substrate Defect Detection Using Hierarchical Classifiers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting defects in electronic substrates, such as printed circuit boards, face challenges in accurately identifying and classifying defects amidst artifacts and variations, leading to potential misidentification and inefficiencies in the manufacturing process.
Innovation Solution
A method involving image processing, defect presence classifiers, and defect type classifiers that utilize techniques like key point detection, registration, subtraction, filtering, and machine learning algorithms, including autoencoders and generative adversarial networks, to generate and classify defect image patches, determining defect location, size, and type, and storing this information in a database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional defect detection methods are used, then the detection process is simple, but the accuracy of defect identification is low due to misidentification among artifacts and variations
Solution Approach 1:
The patent segments the defect detection process into multiple specialized stages: image acquisition, pre-processing, defect candidate identification, artifact filtering, and verification. Each stage handles specific aspects of the detection task, improving overall accuracy while managing complexity through modular organization. The segmentation allows the system to focus computational resources on critical detection steps rather than attempting single-stage detection.
Solution Approach 2:
The patent introduces intermediary components such as reference databases, classification models, and filtering algorithms that act as mediators between raw image data and final defect identification. These intermediaries process and refine the data at each stage, reducing misidentification by comparing against known defect patterns and filtering out artifacts before final classification.
2Loss of information
If comprehensive defect analysis is performed, then detailed defect information is obtained, but the processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing images (alignment, normalization, enhancement) before defect detection and by pre-establishing reference databases of known defects and artifacts. This preliminary preparation reduces the computational burden during actual detection, allowing comprehensive analysis without excessive processing time during production inspection.
Solution Approach 2:
The patent applies local quality by focusing detailed analysis only on regions identified as potential defect candidates, rather than uniformly processing the entire image. Once defect regions are localized, comprehensive analysis is applied specifically to those areas, while other regions receive minimal processing, thus balancing information completeness with processing efficiency.
3Measurement precision
If multiple classification stages are used, then defect classification accuracy is improved, but the system complexity increases
Solution Approach 1:
The patent segments the classification task into multiple hierarchical stages: initial defect candidate identification, artifact filtering, defect type classification, and severity assessment. Each classification stage builds upon previous results, improving overall accuracy while maintaining manageable complexity through progressive refinement rather than attempting all classifications simultaneously.
Solution Approach 2:
The patent uses copying by creating and maintaining reference databases that contain copies of known defect patterns, artifact characteristics, and classification rules. These reference copies are used for comparison during detection, enabling accurate multi-stage classification without requiring the system to re-analyze all characteristics at each stage, thus managing complexity through reusable reference data.
Data Source
AI summary
This disclosure provides systems, methods, and apparatus detecting defects in a substrate. An image of the substrate is compared with a reference image to identify potential defects. Images corresponding to the potential defects are processed sequentially by a set of classifiers to generate a set of images that include a defect. The set of classifiers can be arranged to have increasing accuracy. A subset of the images corresponding to the potential defects is processed by a type classifier that can determine the type, size, and location of the defect in the images. The defects can be further processed to determine the severity of the defects based on the location of the defects on the substrate.