Approximate computing for onboard anomaly detection from hyperspectral images

Abstract

Interest on anomaly detection for hyperspectral images has increasingly grown during the last decades due to the diversity of applications that benefit from this technique. However, the high computational cost inherent to this detection procedure seriously limits its processing efficiency, especially for onboard application scenarios. In this paper, a novel spectral and spatial approximate computing approach, named SSAC is proposed for onboard anomaly detection from hyperspectral images. To efficiently design the proposed approach, two preliminary aspects have been deeply analyzed in this work. First, data correlation in hyperspectral images in both spectral and spatial dimensions has been analyzed. The high data correlation in both spectral and spatial dimensions is considered to be one of the cornerstones of the SSAC approach. Second, the error resilience of a popular hyperspectral anomaly detection algorithm in both data level and algorithm level has been analyzed, which is considered to be another cornerstone of the SSAC approach. Based on the outcomes of this analysis, the processing of spectrally and spatially degraded images has been employed for reducing computation complexity in onboard hyperspectral anomaly detection scenarios in this work. Performance assessment tools such as ROC curves, Cost curves, and computing times have been used for evaluating the computing accuracy and efficiency of our proposal. The results obtained with a nonlinear anomaly detector for hyperspectral imagery, such as the well-known kernel RX-algorithm, show that the proposed SSAC approach greatly improves anomaly detection efficiency compared to the traditional method with negligible degeneration in accuracy. This is an important achievement to meet the restrictions of onboard hyperspectral anomaly detection scenarios.