Comments on "Orthogonal Subspace Projection (OSP) Revisited: A Comprehensive Study and Analysis

A recent paper discussed the orthogonal subspace projection (OSP) algorithm. That paper advocated the use of OSP for target detection and material abundance estimation for mixed pixels in hyperspectral images. However, it was proved in an older paper that an algorithm called the constrained signal detector (CSD) generally outperforms the OSP for the applications of target detection and abundance estimation. The purpose of this correspondence is to review the differences in detection and estimation performance between OSP and CSD