基于DSM排序的真正射影像遮挡区自动检测算法  

Algorithm for Automatic Occlusion Detection of True Orhtophoto Based on Sorted DSM

作　　者：张丽[1,2,3] 汤晓涛[2,3] 莫德林[1] 

机构地区：[1]信息工程大学地理空间信息学院,河南郑州450052 [2]西安测绘研究所,陕西西安710054 [3]地理信息工程国家重点实验室,陕西西安710054

出　　处：《测绘科学与工程》2015年第1期31-36,共6页Geomatics Science and Engineering

摘　　要：遮挡区自动检测是真正射影像制作过程中的关键技术环节。本文提出一种改进的基于DSM排序影像遮挡区自动检测算法，将DSM以地底点为中心分为多块进行排序，并进行了基于GPU的并行加速实现；以分块排序后的DSM为基础进行遮挡区检测，降低了算法的时间复杂度。实验结果表明，提出的方法能有效进行影像遮挡区检测，且执行效率较高。Automatic occlusion detection is a key technology in true orhtophoto generation process. An improved method for automatic occlusion detection based on sorted DSM is proposed in this paper. Setting the ground nadir point as the center, the DSM is divided into several blocks and then all the blocks are sorted and accelerated by GPU - based parallel process. The occlu- sion area is detected based on the divided and sorted DSM, thus the time complexity can be reduced. The experiment results show that this method can conduct occlusion detection effectively with a much higher efficiency.

关 键 词：遮挡区检测 DSM排序 真正射影像 

分 类 号：P231[天文地球—摄影测量与遥感]