視域計算之研究

Abstract

本研究探討視域計算的方法，及布林視域、機率視域、模糊視域之關係。本研究應用直線法逐點法開發視域計算程式，並GIS程式集中之視域功能，ARC/INFO（visibility）、GRASS（r.los），以實例進行比較測試，以熵值等六種不同之指標評估產生之視域圖間的相似程度。經實例分析，發現以誤差模擬的方法計算之機率視域，可提高不同演算法間之相似程度。

Two algorithms for the viewshed computation, the Bensenham method and the fixed interval method, are implemented in this study. Differences between these two algorithms and the software implemented in the GIS packages, namely, ARC/INFO (visibility) and GRASS (r.los), are evaluated with real terrain data sets. Six indices are used to measure the similarity between the viewshed computed from different algorithms. The concept of binary, probable and fuzzy viewshed are also investigated and experimented. It is found that there are apparent differences between different algorithms/software implementations. Comparing with binary viewshed, the probable viewshed computed with added simulated elevation error, provides higher correlation between different algorithms.