Small sample industrial product defect classification method based on two-stage transfer learning

Abstract

A small sample industrial product defect classification method based on two-stage transfer learning comprises the following steps: S1, collecting positive and negative samples to form a data set; S21,expanding the number of the negative samples in the data set by 2-3 times by using an image data enhancement means, randomly selecting positive samples of which the number is equivalent to the numberof the expanded negative samples, and forming a data subset of which the number is balanced; S22, forming another data set subset by using the remaining positive samples; S31, selecting a CNN detection model, and carrying out first-stage training; S32, carrying out training in the second stage on the data set subset containing the remaining positive samples and the expanded negative samples; andS4, after the model training in the step S32 is converged, testing the classification performance of the model on the test set, if the requirements are met, performing online test, otherwise, repeatedly dividing the data subsets and the model training process, and repeating the steps S21 to S32 until the requirements are met. The method has the following beneficial effects: 1, the method has defect image high-dimensional features with better performance; 2, the representation capability of the model on an industrial product image is improved; and 3, the model training strategy has good universality.

Description

technical field [0001] The invention belongs to the technical field of visual inspection, and in particular relates to a small-sample industrial product defect classification method based on two-stage transfer learning. Background technique [0002] With the development of intelligent manufacturing, the demand for production automation of industrial products is becoming more and more urgent, and the requirements for product quality are also getting higher and higher. In the current industrial production practice, most of the production technology process has been automated. However, it is still difficult to automate the product appearance quality inspection and inspection process, which requires a lot of manual inspection, which affects production efficiency and product quality. Therefore, the research on the appearance defect detection technology and system of industrial products is very important. [0003] The basic process of product appearance inspection is to first obt...

AI Technical Summary

Problems solved by technology

Traditional image processing algorithms use the method of feature engineering to achieve classification by calculating artificially designed image features, but this method relies heavily on people, difficult to debug, and poor adaptability

With the development of deep learning technology, the application of defect detection technology based on deep learning model is becoming more and more widespread. However, the number of defect samples in the production process of industrial products is small, the collection is difficult, and the labeling process is time-consuming and labor-intensive. There are few defective samples for training, but the number of normal samples is large. Therefore, the number of defective samples and normal samples is unbalanced. The deep learning model trained on this unbalanced data set will cause the model to be biased towards A large number of normal samples leads to a decrease in the detection accuracy of defective samples, making it difficult to guarantee the pass rate of the final product

Images