FPGA Design of the N-FINDR Algorithm for Spaceborne Hyperspectral Missions

Hyperspectral imaging is a new technique in remote sensing which generates hundreds of images (at different wavelength channels) for the same area on the surface of the Earth. Each pixel collected by a hyperspectral remote sensing instrument is in fact a spectral signature of the underlying materials. Many algorithms attempt to find pure spectral signatures in the image data, called endmembers, and use this information to identify the underlying materials within each pixel. The N-FINDR algorithm is one of the most popular and widely used, despite its high computational complexity when applied to high-dimensional images. However, implementations of this algorithm on reconfigurable hardware are not yet available in the literature, despite the potential advantages that can be gained from processing the hyperspectral images onboard the sensor using specialized hardware devices. In this paper, we present a field programmable gate array design of the algorithm, which has been implemented on a Virtex-4 XC4VFX60 FPGA and evaluated using the well-known “Cuprite” image (a standard benchmark in hyperspectral imaging applications). Our experimental results demonstrate that a hardware version of the N-FINDR algorithm can significantly outperform an equivalent software version and retain excellent pure spectral extraction accuracy.