Reconfigurable acceleration for hyperspectral target detection

The abundant hardware resources, flexibility, scalability and improved performance in modern Field programmable Gate Arrays (FPGA) create a promising solution for on-board high performance, low power computational platforms. They offer a low-cost alternative to Application Specific Integrated Circuit (ASIC) devices. This paper elaborates on developing a target detection system based on Constraint Linear Discriminant Analysis (CLDA) on FPGAs for real-time or near real-time target detection using hyperspectral data. The algorithm and its implementation are described and details of the important subsystem architectures are illustrated.