Abstract:
In this paper, near real-time GPU implementations of two efficient SAR change detection methods using closed-form Kullback-Leibler divergence between generalized Gamma di...Show MoreMetadata
Abstract:
In this paper, near real-time GPU implementations of two efficient SAR change detection methods using closed-form Kullback-Leibler divergence between generalized Gamma distributions (KL-GGD) and two densities approximated by Edgeworth series (KL-EW) are investigated and compared in terms of both accuracy and speed. The near real-time implementations of the proposed methods using Compute Unified Device Architecture (CUDA) on Graphics Processing Units (GPUs) are described and evaluated. The computation time of the parallel implementation on GPU is compared with the C/C++ implementation on Central Processing Unit (CPU). Our experimental results show that the GPU implementation is at least twenty times faster than the CPU implementation.
Date of Conference: 22-27 July 2012
Date Added to IEEE Xplore: 10 November 2012
ISBN Information: