Autoencoder Defect Detection Using Student's t-Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional manual detection of surface defects in industrial products is subjective, inefficient, and costly, lacking real-time performance.
Innovation Solution
A computer device method using an autoencoder model and student's t-distribution to analyze images, training on unblemished images to detect defects by calculating reconstruction errors and filtering noise, determining defect presence based on statistical criteria.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual detection is used, then subjective judgment flexibility is maintained, but detection efficiency is low and labor cost is high
Solution Approach 1:
The patent replaces the manual mechanical inspection system with an automated optical detection system using deep learning algorithms. The autoencoder model processes images to identify surface defects, substituting human inspectors with an automated computational system that achieves higher efficiency and consistency without requiring subjective judgment.
2Measurement precision
If traditional image processing methods are used, then system complexity is low, but defect detection accuracy is insufficient
Solution Approach 1:
The patent transforms the detection approach by changing the fundamental parameters of image analysis. Instead of using traditional fixed thresholding or edge detection parameters, the system employs a trained autoencoder model that learns optimal parameter representations from training data. The model transforms input images into a latent space representation, enabling more accurate defect identification through learned feature patterns rather than fixed mathematical parameters.
3Measurement precision
If autoencoder model with student's t-distribution is used, then defect detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-training the autoencoder model using student's t-distribution loss function during the offline training phase. This preliminary training establishes robust feature extraction capabilities and noise filtering mechanisms before actual detection. During online detection, the pre-trained model processes images efficiently without requiring real-time iterative optimization, thus reducing computational resource consumption during production while maintaining high accuracy.
Data Source
AI summary
A method for detecting defects in images, is employed in a computer device, and stored in a storage medium. The method trains an autoencoder model using unblemished images, inputting an image to be detected into the autoencoder model, and obtaining a reconstructed image. An image error is calculated between the image to be detected and the reconstructed image, and the image error is inputted into a student's t-distribution and a calculation result is obtained. In response that the calculation result falls within a preset defect determination criterion range, the image to be detected is determined to be an unblemished image. In response that the calculation result does not fall within the preset defect determination criterion range, the image to be detected is determined to be a defective image. The method improves the efficiency and accuracy of defect detection.