Application of a two-stage algorithm for adaptive detection in hyperspectral imaging

Adaptive detection has a rich history in the radar community, and a number of other areas have borrowed heavily from constructs developed in this field. The task of target detection in hyperspectral imaging (HSI) is one such area that has recently begun to take advantage of parallels from the field of radar array processing. While there are key differences between the physical setup and data collected by radar and hyperspectral systems, the formulation of the adaptive detection problem is remarkably similar. In this paper we apply a two-stage detection approach originally developed for STAP in airborne radar to adaptive target detection in hyperspectral imaging. Touching first on the component algorithms involved, and then on their similarities and differences, we show that this two-stage approach has an interesting conceptual interpretation for HSI that comes from the multi-dimensional Euclidean geometry of the spectral measurement space. However, due to the nature of hyper-spectral target signatures and background statistics, we propose that there are many scenarios where HSI detection may be better served by reversing the order of stages applied to the data as laid out in previous formulations. Detection results for ground targets are presented to illustrate the potential of this new two-stage approach in hyperspectral adaptive detection