TY - GEN
T1 - Parallel Large-Scale Image Processing for Orthorectification
AU - Im, Changjin
AU - Jeong, Jae Heon
AU - Jeong, Chang Sung
N1 - Funding Information:
ACKNOWLEDGMENT This research was supported by the utilization of satellite information project through the Korea Aerospace Research Institute (KARI), Basic Science Research Program through the National Research Foundation f Korea (NRF) funded by the Ministry of Education (2017R1D1A1B03035461), the Brain Korea 21 Plus Project in 2018, and the Institute for Information&Communications Technology Promotion(IITP) grant funded by the Korean government (MSIP) (No. 2018-0-00739, Deep learning-based natural language contents evaluation technology for detecting fake news).
Publisher Copyright:
© 2018 IEEE.
PY - 2019/2/22
Y1 - 2019/2/22
N2 - Recently, as the amount of data that needs to be processed is getting larger, various techniques of parallel processing have been developed to deal with large-scale data efficiently and quickly. In this paper, we shall present fast parallel orthorectification algorithms for large-scale image data on various environments, and compare their performance in terms of speed-up and execution time. Our research consists of two parts: First, we implement parallel orthorectification algorithm on CPU multicore, Xeon-phi multicore and GPU. Second, we analyze these experiment results by comparing the performance of each algorithm. More specifically, we compare the performance of CPU multicore with Xeon-phi and GPU parallelization to find which one is more efficient as well as the performance of Xeon-phi multicore and GPU parallelization. We shall show that for the former, GPU parallelization is more efficient technique than CPU multicore parallelization, while for the latter, Xeon-phi multicore parallelization is more efficient technique than GPU parallelization. This is due to the data upload/download time on GPU. Even if we extend the experiment to infinite-scale, data upload/download time on GPU is still needed. Therefore, Xeon-phi multicore parallelization is better than GPU parallelization not only on the extended environment but also the infinite-scale environment.
AB - Recently, as the amount of data that needs to be processed is getting larger, various techniques of parallel processing have been developed to deal with large-scale data efficiently and quickly. In this paper, we shall present fast parallel orthorectification algorithms for large-scale image data on various environments, and compare their performance in terms of speed-up and execution time. Our research consists of two parts: First, we implement parallel orthorectification algorithm on CPU multicore, Xeon-phi multicore and GPU. Second, we analyze these experiment results by comparing the performance of each algorithm. More specifically, we compare the performance of CPU multicore with Xeon-phi and GPU parallelization to find which one is more efficient as well as the performance of Xeon-phi multicore and GPU parallelization. We shall show that for the former, GPU parallelization is more efficient technique than CPU multicore parallelization, while for the latter, Xeon-phi multicore parallelization is more efficient technique than GPU parallelization. This is due to the data upload/download time on GPU. Even if we extend the experiment to infinite-scale, data upload/download time on GPU is still needed. Therefore, Xeon-phi multicore parallelization is better than GPU parallelization not only on the extended environment but also the infinite-scale environment.
KW - CUDA
KW - large-scale image
KW - multi-GPU
KW - multi-core
KW - openmp
KW - orthorectification
KW - parallel
KW - xeon-phi
UR - http://www.scopus.com/inward/record.url?scp=85063237254&partnerID=8YFLogxK
U2 - 10.1109/TENCON.2018.8650289
DO - 10.1109/TENCON.2018.8650289
M3 - Conference contribution
AN - SCOPUS:85063237254
T3 - IEEE Region 10 Annual International Conference, Proceedings/TENCON
SP - 2153
EP - 2157
BT - Proceedings of TENCON 2018 - 2018 IEEE Region 10 Conference
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 IEEE Region 10 Conference, TENCON 2018
Y2 - 28 October 2018 through 31 October 2018
ER -