Automated Defect Detection Framework Using Classification and Locating Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The precision manufacturing industry faces challenges with manual inspection for product defects, leading to inconsistent quality, high labor demands, and susceptibility to human error, necessitating an automated solution for stable and consistent defect detection.
Innovation Solution
A product defect detection method and system utilizing a defect detection framework comprising a classification network, a locating detection network, and a judgment network, trained on sample images to classify and locate defects in product images, enabling accurate two-stage detection of defect types and positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual inspection is used for product defect detection, then human operators can identify defects, but inspection quality fluctuates due to fatigue and human error
Solution Approach 1:
The patent replaces the mechanical human inspection system with an automated computer vision system comprising image acquisition devices, preprocessing modules, and defect detection networks. This substitution eliminates human fatigue and inconsistency while maintaining high detection accuracy through algorithmic processing of product images.
Solution Approach 2:
The defect detection system performs self-service by automatically acquiring product images, preprocessing them, detecting defects through trained networks, and generating results without human intervention. The system serves itself to complete the entire inspection workflow autonomously, ensuring consistent quality without human operators.
2Productivity
If manual inspection is used for product defect detection, then defects can be identified, but labor demand is high causing workforce shortages
Solution Approach 1:
The patent replaces numerous human inspectors with a single automated computer vision system that can process multiple product images simultaneously. The system uses parallel processing capabilities to handle high volumes of inspection data, dramatically improving productivity while reducing the quantity of human resources required.
3Reliability
If automated inspection is implemented, then consistent defect detection is achieved, but system complexity increases
Solution Approach 1:
The patent segments the defect detection system into distinct functional modules: image acquisition devices for capturing product images, preprocessing modules for image enhancement and preparation, and defect detection networks for analyzing processed images. This segmentation allows each module to be optimized independently while maintaining overall system consistency and reliability.
Solution Approach 2:
The defect detection framework is designed with universal components that can handle multiple defect types and product variations. The trained detection networks can identify various defect categories (surface defects, structural defects, dimensional deviations) using the same underlying architecture, reducing system complexity through multi-functionality.
4Measurement precision
If comprehensive defect detection is performed, then all defect types are identified, but detection time increases
Solution Approach 1:
The patent applies preliminary action through the preprocessing module that prepares images before defect detection by enhancing relevant features, adjusting contrast, and filtering noise. This preliminary preparation enables the detection networks to quickly and accurately identify all defect types without increasing inspection time, as the hard work of image preparation is done in advance.
Solution Approach 2:
The detection system segments defect analysis into specialized networks trained for different defect types (surface defects, structural defects, dimensional deviations). This segmentation allows parallel processing of multiple defect categories simultaneously, maintaining high detection accuracy while reducing total inspection time compared to sequential analysis.
Data Source
AI summary
A product defect detection method, device and system are disclosed. The product defect detection method comprises: constructing a defect detection framework including a classification network, a locating detection network and a judgment network; training the classification network by using a sample image of a product containing different defect types to obtain a classification network capable of classifying the defect types existing in the sample image; training the locating detection network by using a sample image of a product containing different defect types to obtain a locating detection network capable of locating a position of each type of defect in the sample image; inputting an acquired product image into the defect detection framework, inputting a classification result and a detection result obtained into the judgment network to judge whether the product has a defect, and detecting a defect type and a defect position when the product has a defect.