Target detection in hyperspectral images using basic thresholding classifier

TOKSÖZ, Mehmet Altan
Ulusoy, İlkay
In this letter, we propose target detector version of recently introduced basic thresholding classifier for hyperspectral images. The proposed technique is a sparsity-based low complexity detector which achieves high detection rates with very low false alarm rates and performs extremely rapidly. We also propose a new decision metric, background model, and spatial smoothing procedure in order to increase the detection probability further. Experiments show that the presented method outperforms the other state-of-the-art sparsity-based approaches.