A Weighted Spatial-Spectral Kernel RX Algorithm and Efficient Implementation on GPUs.

摘要: The kernel RX (KRX) detector proposed by Kwon and Nasrabadi exploits a function to obtain better detection performance. However, it still has two limits that can be improved. On the one hand, reasonable integration of spatial-spectral information used further improve its accuracy. other parallel computing reduce processing time in available KRX detectors. Accordingly, this paper presents novel weighted (WSSKRX) implementation on graphics units (GPUs). WSSKRX utilizes spatial neighborhood resources reconstruct testing pixels introducing spectral factor window, thereby effectively reducing interference background noise. Then, is redesigned as mapping trick implement anomaly detection. In addition, powerful architecture based GPU technique designed accelerate WSSKRX. To substantiate performance algorithm, both synthetic real data are conducted for experiments.