A comparative study of subpixel target detectors for hyperspectral imagery

Target detection is one of the most challenging issues of remotely sensed data. Due to the limited ground sampling distance (GSD), targets of interest occur at subpixel level and mix with more signatures from the background. In such cases, spatial characteristics of targets is hard to acquire and the only way to overcome such problem is to take advantage of spectral information. Based on the spectral characteristics of ackground and the targets to be detected, several methods have been proposed. Some of these methods assume a physics-based approach, while the other may be purely statistical. Recently, some detectors which take into account the two above mentioned approaches (i.e. hybrid detectors) have been developed as well. This paper presents a clear understanding of 10 subpixel target detectors and also provides a performance comparison among these methods under different circumstances

Keywords:

Target detection , hyperspectral data and spectral mixture models