Cross-Category Defect Detection Using Paired Product Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing defect detection methods in industrial products are limited to a single category and require re-training for different product categories, hindering production efficiency and generalization.
Innovation Solution
A defect detection architecture using a large model that integrates feature extraction and up-sampling to enable cross-category defect detection, allowing for defect-free and to-be-detected product image analysis without re-training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a feature library is used for defect detection, then defect detection can be performed on a single category of products, but the model needs to be re-trained when detecting another category of products
Solution Approach 1:
The patent applies universality by designing a defect detection model that can handle multiple product categories simultaneously. The model integrates feature extraction, feature library matching, and defect detection into a unified framework that processes different product types without requiring separate models or re-training, thereby achieving cross-category defect detection capability while maintaining detection accuracy
Solution Approach 2:
The patent segments the defect detection process into distinct functional modules: feature extraction module, feature library module, and defect detection module. This segmentation allows each module to be optimized independently while working together in a coordinated manner, enabling the system to handle diverse product categories through a modular architecture that can be configured for different products without complete re-training
2Reliability
If a feature library is used for defect detection, then defect detection can be performed, but production efficiency is hindered due to re-training requirements
Solution Approach 1:
The patent implements preliminary action by pre-building a comprehensive feature library that contains features from multiple product categories before actual defect detection begins. This pre-prepared feature library enables the model to immediately detect defects in various product categories without requiring time-consuming re-training processes, thereby maintaining reliable defect detection capability while significantly improving production efficiency
Data Source
AI summary
The present disclosure discloses an industrial product defect detection method and apparatus, a device, and a medium, and belongs to the field of image processing, and may be applied to various scenarios such as a cloud technology, artificial intelligence, intelligent transportation, and assisted driving. One method includes: obtaining a first product image and a second product image; separately performing feature extraction on the first product image and the second product image to obtain a first image feature and a second image feature; merging the first image feature with the second image feature to obtain a first intermediate feature; inputting the first intermediate feature into a defect detection model to obtain an inference feature; performing up-sampling on the inference feature to obtain a second intermediate feature; and obtaining information about a position of a defect based on the second intermediate feature.