摘要
针对通用计算平台下SAR成像算法效率低下的问题,提出了一种基于CUDA的SAR成像算法并行化实现方法。在分析CUDA工作原理及CS算法并行性特征的基础上,详细描述了算法每个步骤的CUDA实现。实验结果表明了该算法的高效性,优化后的CS算法提速比达到了10~20倍。
A parallel implementing method for synthetic aperture radar (SAP.) imaging algorithm based on compute unified device architecture (CODA) is proposed in order to solve the low efficiency of SAR. imaging algorithm on general computing platform. CUDA principle and chirp scaling (CS) algorithm parallel feature are analyzed. On the basis of this, the CUDA implementation of each algorithm step is described in detail. Experimental results show that the optimized CS algorithm is 10-20 times faster. Therefore, the algorithm is effective.
出处
《计算机与网络》
2012年第7期55-57,共3页
Computer & Network