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

Chunhui Zhao,Jiawei Li,Xifeng Yao,Meiling Meng

doi:10.3390/s17030441

Abstract

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

Highlights

Hyperspectral imagery (HSI) is served as a three-dimensional cube which contains of two spatial dimensions and one spectral dimension
The parallel version is implemented in Compute Unified Device Architecture (CUDA) C programing language for graphics processing units (GPUs) cards performed a 2.0
We evaluate the parallel performance of weighted spatial-spectral kernel RX (WSSKRX) in in

Summary

Introduction

Hyperspectral imagery (HSI) is served as a three-dimensional cube which contains of two spatial dimensions and one spectral dimension. It provides the ability to distinguish the differences of ground-object spectra, so it has a wide range of applications in target detection [1]. Based on the availability of the prior information, the target detection algorithms can be divided into unsupervised and supervised ones. As accurate prior knowledge is difficult to obtain, unsupervised algorithms (anomaly detection algorithms) have drawn wide interest. Detection uses the differences between targets and the backgrounds to detect anomalies [2,3,4,5,6,7]. A widely used anomaly detection algorithm is the RX detector proposed by Reed and Yu [8]

Results

Discussion

Conclusion