Efficient Implementation of a Fast Viewshed Algorithm on SIMD Architectures

View shed refers to the land area that is visible to an observer placed in a point of a terrain. Due to the advances in remote sensing technologies the volume of data is today beyond the capability of traditional GIS tools and therefore new and fast algorithms become essential. In this paper we present an efficient implementation of the XDRAW algorithm [5] to quickly compute view sheds on very large digital elevation models. We redesign the algorithm to make it IO-efficient and compatible with modern SIMD architectures. Our implementation is able to compute view sheds on digital elevation models at the rate of 109 points per second on an Intel quad-core CPU with AVX2 technology, which makes the algorithm suitable for real-time applications.