Abstract: Visibility analysis has been widely used in GIS applications. Viewshed and LOS are its two elementary factors. There have been many methods for viewshed computing using different algorithms and programming techniques, which lead to big differences in consumed time and accuracy. In order to solve the problem that due to the memory limit massive terrain data are unable to be stored and processed, this paper presents an efficient algorithm to compute viewshed on terrain stored in external memory. Through the analysis of the algorithm and the experimental verification, it shows that the algorithm is superior to other existing algorithms.

Key words: Geographic Information System（GIS）, terrain model, external memory processing, viewshed, visual map

摘要： 可视性分析广泛应用于GIS的各个方面，视线和视域是其最基本的两个因子。视域的计算方法有很多种，在计算时因为采用的算法和实现算法所运用的编程方法不同，导致效率和精度有很大差异。为了解决由内存受限而无法存储和处理海量区域地理信息数据的问题，提出一种在外部存储器上计算地形视域的高效算法，对该算法进行了分析，并进行了实验验证，表明该算法明显优于现有的其他算法。

关键词: 地理信息系统（GIS）, 地形模型, 外部存储处理, 视域, 可视性地图