Intelligent metallographic detection rating method and system based on deep learning

Abstract

The invention provides an intelligent metallographic detection rating method and system based on deep learning. The method comprises the following steps: collecting a metal sample image; making improvement on the basis of a U-net full convolutional neural network to obtain the construction of the improved full convolutional neural network; automatically segmenting the acquired metal sample image through an improved full convolutional neural network to obtain a metallographic segmentation image; performing automatic rating classification on the obtained metallographic segmentation map through adeep neural network. According to the method, the deep learning algorithm is adopted, the improvement is carried out on the basis of the U-net full convolutional neural network, the complexity and instability of traditional image manual feature extraction are avoided, and meanwhile, the applicability of feature extraction is improved based on parameter optimization of a gradient descent method; the method is based on the deep neural network, so the system calculation time is greatly improved, and tedious classification calculation is avoided; the method is high in segmentation and classification accuracy, segmentation and grading are fused through one key, multi-step operation of a traditional method is avoided, and the method is flexible and convenient.

Description

technical field [0001] The invention belongs to the technical field of visual measurement and detection, and in particular relates to an intelligent metallographic detection and rating method and system based on deep learning. Background technique [0002] In modern industrial society, the quality requirements for metal raw materials are getting higher and higher in all walks of life. Poor grades of raw materials will bring various safety hazards, such as broken shafts of common cars, corrosion and rust of metal products, etc. In order to For the classification of the grade of metal raw materials, the metal raw materials will be sampled before leaving the factory, and then graded. Grain size is an important criterion for judging. The more metal particles per unit area, the better the performance of the metal. Traditional artificial Grading is interfered by various factors, so it is impossible to quickly and accurately grade metal raw materials. With the development of digita...

AI Technical Summary

Problems solved by technology

[0002] In modern industrial society, the quality requirements for metal raw materials are getting higher and higher in all walks of life. Poor grades of raw materials will bring various safety hazards, such as broken shafts of common cars, corrosion and rust of metal products, etc. In order to For the classification of the grade of metal raw materials, the metal raw materials will be sampled before leaving the factory, and then rated. Grain size is an important criterion for judging. The more metal particles per unit area, the better the performance of the metal. Traditional artificial The rating is interfered by various factors, and it is impossible to quickly and accurately rate metal raw materials. With the development of digital image processing technology, image processing technology is more and more used in metallographic rating.

[0003] At present, there are few equipment and systems for metallographic grading on the market, and most of them need to do complex preprocessing on the original metallographic image, then manually select features, and then classify and grade, which is not only time-consuming, but also the accuracy rate cannot be obtained. Guarantee, and most of them can only be rated for specific products, not universal and transplantable

Images