无人机影像SIFT特征匹配的集群并行处理方法  

Cluster Parallel Processing Method of UAV Images Matching Based on SIFT Features

作　　者：李劲澎[1] 龚志辉[1] 张婷[2] 

机构地区：[1]信息工程大学测绘学院,河南郑州450052 [2]河北省土地整理服务中心,河北石家庄050051

出　　处：《测绘科学技术学报》2012年第6期440-444,共5页Journal of Geomatics Science and Technology

基　　金：国家自然科学基金项目(41101396)

摘　　要：简要介绍了国内外摄影测量并行处理技术发展的情况。在分析SIFT(尺度不变特征变换)算法和集群并行算法特点的基础上,设计了一种无人机影像匹配并行处理的方法;兼顾计算局部化和负载均衡提出了分层列划分的数据划分策略;并通过重叠通信和计算来减少通信开销,借助时间分布图对算法进行了量化分析。实验证明该方法可扩展性较好,适合在集群平台上进行影像匹配高效处理。The present research progress of parallel processing in photogrammetry was briefly introduced. Based on the analysis of the SIFT（Scale Invariant Feature Transform） algorithm and the character of parallel algo- rithm in cluster environment, a parallel processing method for UAV（unmanned aerial vehicle） serial images matching was designed. A striped partitioning strategy based on delamination was proposed, in which both localization of computing and load balance were taken account of. To reduce the cost of communication, the technique that overlaps data communication and computing was used. Then, a quantitative algorithm analysis was given with time distributing diagram. The experimental results showed that the method had good scalability and compatibility to processing the images matching efficiently in cluster environment.

关 键 词：尺度不变特征变换 集群 通信 负载均衡 并行处理 

分 类 号：P231[天文地球—摄影测量与遥感]