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

View Article: PubMed Central - PubMed

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.

No MeSH data available.


The grayscale detection results of GRX, LRX, LKRX, HKRX and WSSKRX on three datasets synthetic dataset San Diego Airport dataset SpecTIR dataset. (a) The grayscale of synthetic dataset; (b) The grayscale of San Diego Airport dataset; (c) The grayscale of SpecTIR dataset.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5375727&req=5

sensors-17-00441-f013: The grayscale detection results of GRX, LRX, LKRX, HKRX and WSSKRX on three datasets synthetic dataset San Diego Airport dataset SpecTIR dataset. (a) The grayscale of synthetic dataset; (b) The grayscale of San Diego Airport dataset; (c) The grayscale of SpecTIR dataset.

Mentions: Figure 13 and Figure 14 show the grayscale and three-dimensional plots outputs of GRX, LRX, LKRX, HKRX and WSSKRX on three hyperspectral images, respectively. In Figure 13, LKRX, HKRX and WSSKRX can detect more abnormal pixels compared with GRX and LRX. This is because LKRX, HKRX and WSSKRX which are based on the kernel function exploit the nonlinear characteristics between spectral bands. Simultaneously, as WSSKRX algorithm effectively combines the spatial-spectral information, the detection accuracy is higher than other four detectors. In Figure 14, it can be seen that WSSKRX has better ability of suppressing noise interference compared with the LKRX and HKRX. From the observations, the separability performance of WSSKRX is better than other algorithms.


A Weighted Spatial-Spectral Kernel RX Algorithm and Efficient Implementation on GPUs
The grayscale detection results of GRX, LRX, LKRX, HKRX and WSSKRX on three datasets synthetic dataset San Diego Airport dataset SpecTIR dataset. (a) The grayscale of synthetic dataset; (b) The grayscale of San Diego Airport dataset; (c) The grayscale of SpecTIR dataset.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5375727&req=5

sensors-17-00441-f013: The grayscale detection results of GRX, LRX, LKRX, HKRX and WSSKRX on three datasets synthetic dataset San Diego Airport dataset SpecTIR dataset. (a) The grayscale of synthetic dataset; (b) The grayscale of San Diego Airport dataset; (c) The grayscale of SpecTIR dataset.
Mentions: Figure 13 and Figure 14 show the grayscale and three-dimensional plots outputs of GRX, LRX, LKRX, HKRX and WSSKRX on three hyperspectral images, respectively. In Figure 13, LKRX, HKRX and WSSKRX can detect more abnormal pixels compared with GRX and LRX. This is because LKRX, HKRX and WSSKRX which are based on the kernel function exploit the nonlinear characteristics between spectral bands. Simultaneously, as WSSKRX algorithm effectively combines the spatial-spectral information, the detection accuracy is higher than other four detectors. In Figure 14, it can be seen that WSSKRX has better ability of suppressing noise interference compared with the LKRX and HKRX. From the observations, the separability performance of WSSKRX is better than other algorithms.

View Article: PubMed Central - PubMed

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.

No MeSH data available.