Microstructure image recognition method and device based on self-supervised neural network

Abstract

The invention discloses a microstructure image recognition method and device based on a self-supervised neural network. The method comprises the following steps: pre-training a convolutional network; performing batch feature extraction on the preprocessed standard microstructure chart by using a pre-trained convolutional network to obtain a feature vector set, and storing the feature vector set into a database B with an HDF5 feature vector storage structure; using a pre-trained convolutional network to perform batch feature extraction on the preprocessed to-be-recognized microstructure chart to obtain a feature vector V; calculating the similarity between the feature vector V and all feature vectors in the database B, performing ascending sorting according to the similarity to obtain k feature vectors closest to the feature vector V, and mapping the k feature vectors to the database A to obtain corresponding k standard microstructure diagrams; and counting the specific names of the microstructures to which most of the k standard microstructure diagrams belong, and determining the specific names as the specific names of the microstructure diagrams to be recognized. According to the invention, the recognition accuracy and recognition speed of the microstructure image are improved.

Description

technical field [0001] The invention relates to the technical field of recognition of microstructure images, in particular to a method and device for recognizing microstructure images based on a self-supervised neural network. Background technique [0002] The current method of microstructure image recognition generally uses traditional digital image processing algorithms to perform batch preprocessing and feature extraction on all standard microstructure images in database A, and store all the extracted feature vectors in database B. . For the microstructure diagram to be identified requested by the user, use the same feature extraction algorithm to extract the feature vector, and then calculate the similarity between the feature vector and all the feature vectors in database B and sort them in ascending order, so as to obtain the closest microstructure to be identified. K eigenvectors of the eigenvectors of the structural diagrams (the eigenvectors of the standard microst...

AI Technical Summary

Problems solved by technology

[0004] (1) The feature extraction work of microstructure images is inefficient, and the discriminative information of the original image contained in the extracted features is very limited, which affects the recognition accuracy of the system;

[0005] (2) The database used to store feature vectors is a relational database, which cannot organize feature vector sets in a targeted manner to achieve fast access

Images