Dense Computing Task Analysis of Multi-view Matching Method and GPU Implementation
For the parallel processing of multi-view matching method of high resolution digital imagery,dense computing tasks and crucial parallel computing aspects were analyzed.And then,based on GPU implementation scheme of single stereo matching method,an improved GPU implementation scheme was proposed.Based on practical image datasets,experiments were done and the experimental results proved that the acceleration rate of the improved GPU implementation scheme could reach 20 to 100 for dense computing tasks,and compared with single CPU processing,possessed greater processing superiority.
Multi-view Matching Task GPU Parallel Acceleration Rate
Chenguang Dai Song Ji Jingyu Yang
State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan Univ Key Laboratory of Mapping from Space National Administration of Surveying, Mapping and Geoinformatio
国际会议
厦门
英文
1-6
2012-12-16（万方平台首次上网日期，不代表论文的发表时间）