Performance improvement of viewshed analysis using GPU

Abstract

In this paper, the implementation of viewshed analysis on many-core GPU (Graphic Processing Unit) is presented. The implementation is based on DEM elevation data and CUDA programming framework to access massively parallel GPU architecture for general purpose computing. Experimental evaluation indicates improvement in performance in respect to CPU-based solution and shows feasibility of using GPU and CUDA for viewshed analysis and similar Geographic Information Systems (GIS) algorithms.