Non-imaging target recognition algorithm based on projection matrix and image Euclidean distance by computational ghost imaging

Abstract

With the development of target recognition technology, people pay more and more attention to target recognition. This paper proposed a non-imaging target recognition technology based on projection matrix and image Euclidean distance by computational ghost imaging. The main features of spatial sample images are extracted by principal component analysis. The projection matrix of eigenspace is used as the modulation matrix in computational ghost imaging algorithm to reduce data redundancy. And by calculating the image Euclidean distance, the correlation of pixels is taken into full consideration in spatial position, so as to improve the accuracy of traditional target recognition algorithms. In this paper, the target recognition is realized by directly using the detector observations of the computational ghost imaging system based on the eigenspace projection matrix and the calculation of the image Euclidean distance. The algorithm not only has better recognition performance without reconstructing the image.