Models and Algorithms of Hyperspectral Image Anomaly Detection Based on Low-Rank Technology

Recently, the anomaly detection (AD) problem of hyperspectral image has attracted a lot of attention due to its wide application in many fields. In this paper, the basic theory of hyperspectral image anomaly detection algorithm based on lowrank technology is introduced, several models by means of lowrank decomposition technology and sparse representation technology are studied, which include the robust principal component analysis (RPCA) model and its improved model, lowrank representation (LRR) model and low-rank representation model of spectral unmixing technology. Then, as for the optimization problems of different models, the augmented Lagrange multiplier method (ALMM), alternating direction method (ADM) and their linearization algorithms are analyzed. Finally, the conclusion of this paper and prospect of the research are given.