Title :
Energy-efficient high-performance SAR image geocoding with NVIDIA CARMA and its application in stereo radargrammetry
Author :
Balz, Timo ; Lu Zhang ; Mingsheng Liao
Author_Institution :
State Key Lab. of Inf. Eng. in Surveying, Wuhan Univ., Wuhan, China
Abstract :
Energy is the major limiting factor for high-performance computing. Nowadays, modern Graphics Processing Units (GPU) use less power per floating-point operation. With the CUDA on ARM architecture, a low-power high performance architecture is available and we present first results on using this architecture for fast SAR geocoding and demonstrate the use of fast SAR geo-coding on GPU for a flexible SAR stereo-radargrammetry approach. We demonstrate the speed-up of SAR geo-coding with GPU implementations on an NVIDIA Tesla C2070 and the CARMA DevKit.
Keywords :
geophysical image processing; graphics processing units; image coding; parallel architectures; radar imaging; stereo image processing; synthetic aperture radar; ARM architecture; CARMA DevKit; CUDA; GPU; NVIDIA Tesla C2070; energy-efficient high-performance SAR image geocoding; flexible SAR stereo-radargrammetry approach; floating-point operation; graphics processing unit; Abstracts; Graphics processing units; Indexes; Logic gates; GPU; SAR; geo-coding; radargrammetry;
Conference_Titel :
Geoscience and Remote Sensing Symposium (IGARSS), 2013 IEEE International
Conference_Location :
Melbourne, VIC
Print_ISBN :
978-1-4799-1114-1
DOI :
10.1109/IGARSS.2013.6723204
