Recursive RX with Extended Multi-Attribute Profiles for Hyperspectral Anomaly Detection

Hyperspectral anomaly detection (HAD) plays an important role in military and civilian applications has attracted a lot of research. The well-known Reed–Xiaoli (RX) algorithm is the benchmark HAD methods. Based on RX model, many variants have been developed. However, most them ignore spatial characteristics hyperspectral images (HSIs). In this paper, we combine extended multi-attribute profiles (EMAP) to propose Recursive with Extended Multi-Attribute Profiles (RRXEMAP) algorithm. Firstly, EMAP utilized extract structure information HSI. Then, simple method background purification proposed. That is, purified by utilizing detector remove pixels that are more likely be anomalies, which helps improve ability estimation. addition, parameter control level can selected experiments. Finally, used again between feature new distribution judge anomaly. Experimental results six real datasets synthetic dataset demonstrate effectiveness proposed RRXEMAP importance using purity means. Especially, abu-airport-2 dataset, AUC value obtained present 0.9858, higher than second one, CRD, 0.0198.