Power image classification algorithm based on improved collaborative representation

Abstract

As a typical linear representation method, collaborative representation has become an important research direction in the field of power image classification. Traditional cooperative representation algorithms often ignore the competitiveness and distinguish ability of each kind of samples, which affects the performance of power image classification. In order to further improve the accuracy of power equipment image recognition, this paper proposes an image classification algorithm based on improved cooperative representation, which makes full use of the competition between each kind of samples and the local geometric structure characteristics of samples. Experiments on power image data sets with and without noise show that the proposed algorithm has good classification performance.