Realistic terrain features and the complexity of joint viewsheds

Abstract

Computing viewsheds from different viewpoints is an important procedure with many applications in multiple Geographic Information Science (GIS) fields. While much research has been done on viewsheds obtained from a single viewpoint, viewsheds from multiple viewpoints are still mostly unexplored. This thesis attempts to give more insight into the complexity of multiple viewsheds by analyzing several measures from different GIS fields on real-world terrains. Sky visibility index, terrain ruggedness index, terrain shape index, fractal Dimension, and prickliness were calculated on datasets of around 50 real-world terrains and statistically analyzed with viewsheds generated from multiple viewpoint configurations. Because GIS fields have different preferences on terrain representations, both digital elevation model (DEM) and triangular irregular network (TIN) terrains were used. This thesis shows some relevant insight into the behavior of viewsheds on realistic terrains. It also provides evidence that measures like prickliness are a good indicator for (multiple) viewshed complexity in some common use cases.