State Machine Implementation of ATGP Algorithm F or Target Detection In Hyperspectral Images

In the realm of remote sensing, hyperspectral imaging is a new field of study. Hyperspectral sensors use the surface of the earth to record hundreds of continuous spectral bands as a three-dimensional cube. The maximum data rate of the communication system might be exceeded by the transmission of this massive amount of data to the station on Earth. The ground processing system might also be burdened by the high dimensional data. FPGAs are light, small in size and have parallel processing systems with low power consumption. They have a reconfigurable capability and can also handle ionizing radiation in space. This work's primary contribution is the development and application of a productive, unsupervised method for target detection in hyperspectral image. The classical Orthogonal Subspace Projection (OSP) algorithm is difficult to implement in hardware as it involves matrix inversion operations and also does not meet the real-time requirements. To overcome this limitation, a modified version of ATGP algorithm is adopted with Gram-Schmidt Orthogonalization is implemented in State machine which uses only inner products. This may reduce the complexity of hardware implementation, fasten the processing speed and reduces the error probability. The algorithm is proposed and is implemented in HDL(Hardware Description Language).