Quantitative statistical characterization method of aluminum alloy micron-scale second phase based on deep learning

Abstract

The invention discloses a quantitative statistical characterization method for micron-level second phases of aluminum alloys based on deep learning, comprising: obtaining a feature database of standard samples, using an image segmentation network U-Net based on deep learning to train the feature database, and obtaining U-Net segmentation model, select the parameters corresponding to the optimal precision to establish the U-Net target model; cut the image of the aluminum alloy to be tested and input it into the U-Net target model, and obtain the size, area, Position information, combined with mathematical statistical methods to perform statistical distribution representation of the data set, restore the position information in the test image to the surface of the aluminum alloy to be tested, and obtain the quantitative statistical distribution and visualization results of the entire field of view. Based on the deep learning image segmentation algorithm, the invention automatically recognizes and extracts the micron-level second phase of aluminum alloy, locates and counts the extracted features, and has the characteristics of large field of view, complete information, accuracy and reliability.

Description

technical field [0001] The invention relates to the technical field of characterization of micron-level second phases in aluminum alloys, in particular to a method for quantitative statistical characterization of micron-level second phases of aluminum alloys based on deep learning. Background technique [0002] Unmelted or refractory second-phase particles in aluminum alloys are generally produced during casting and cannot be melted back during subsequent heat treatment and thermal deformation. These refractory second phases break up and elongate upon deformation, aligning linearly in the direction of deformation, consisting of short complementary connected strips. These particles are hard and brittle, and are distributed inside the grain or on the grain boundary. During plastic deformation, pores are easily formed on the phase interface, resulting in micro-cracks, which significantly reduces the fracture toughness of the material. In addition, localized corrosion such as p...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a quantitative statistical characterization method for micron-level second phases of aluminum alloys based on deep learning, which uses high-throughput image data acquisition methods and deep learning algorithms to realize automatic identification of micron-level second phases of aluminum alloys, and combines The mathematical method excavates a variety of characterization parameters of the second phase, quantitatively counts the differences in the full field of view of the material surface and the distribution of partitions, and solves the problem of single quantitative characterization parameters of the second phase in traditional aluminum alloys. It has a large field of view, complete information, and accuracy. reliable features

Images