Nonlinear joint fractional Fourier transform correlation for target detection in hyperspectral image

Abstract

In this paper, we present a spectral joint fractional Fourier transform correlation (SJFRTC) for detecting and identifying very small target with only a few pixels using reflectance spectral information in the hyperspectral imagery. In this technique, fractional correlation (FC) between input spectral signature and known reference spectral signature is performed digitally or optically. Binary differential technique is utilized to enhance the performance of the SJFRTC, resulting in delta-function-like correlation peaks without dc interference. Distinctive and distinguishable correlation performance is obtained using peak-to-clutter ratio (PCR) instead of absolute correlation peak value, which makes the proposed technique independent from the intensity variation of the spectral signature. Simulation results using real-life hyperspectral images show that the binary SJFRTC can detect both single or/and multiple targets successfully with excellent detection capability, as compared to the other traditional hyperspectral detection algorithms.