DocumentCode :
2309394
Title :
GPU-based Parallel Implementation of SAR Imaging
Author :
Xingxing Jin ; Seok-Bum Ko
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Saskatchewan, Saskatoon, SK, Canada
fYear :
2012
fDate :
19-22 Dec. 2012
Firstpage :
125
Lastpage :
129
Abstract :
Synthetic Aperture Radar (SAR) is an all-weather remote sensing technology and occupies a great position in disaster observation and geological mapping. The main challenge for SAR processing is the huge volume of raw data, which demands tremendous computation. This limits the utilization of SAR, especially for real-time applications. On the other hand, recent developments in Graphics Processing Unit (GPU) technology, which obtain general processing capability, high parallel computation performance, and ultra wide memory bandwidth, offer a novel method for computationally intensive applications. This work proposes a parallel implementation of SAR imaging on GPU via Compute Unified Device Architecture (CUDA), and provides a potential solution for SAR real-time processing. The results show that the proposed method obtained a speedup of 31.72, compared to a CPU platform.
Keywords :
graphics processing units; parallel architectures; radar computing; radar imaging; remote sensing by radar; sensors; synthetic aperture radar; CUDA; GPU-based parallel implementation; SAR image processing; all-weather remote sensing technology; compute unified device architecture; disaster observation; geological mapping; graphics processing unit; parallel computation performance; synthetic aperture radar; ultrawide memory bandwidth; CUDA; GPU; Parallel computation; SAR;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Electronic System Design (ISED), 2012 International Symposium on
Conference_Location :
Kolkata
Print_ISBN :
978-1-4673-4704-4
Type :
conf
DOI :
10.1109/ISED.2012.35
Filename :
6526567
Link To Document :
بازگشت