Personalising the viewshed: Visibility analysis from the human perspective

Abstract

Viewshed analysis remains one of the most popular GIS tools for assessing visibility, despite the recognition of several limitations when quantifying visibility from a human perspective. The visual significance of terrain is heavily influenced by the vertical dimension (i.e. slope, aspect and elevation) and distance from the observer, neither of which are adjusted for in standard viewshed analyses. Based on these limitations, this study aimed to develop a methodology which extends the standard viewshed to represent visible landscape as more realistically perceived by a human, called the ‘Vertical Visibility Index’ (VVI). This method was intended to overcome the primary limitations of the standard viewshed by calculating the vertical degrees of visibility between the eye-level of a human and the top and bottom point of each visible cell in a viewshed. Next, the validity of the VVI was assessed using two comparison methods: 1) the known proportion of vegetation visible as assessed through imagery for 10 locations; and 2) standard viewshed analysis for 50 viewpoints in an urban setting. While positive, significant correlations were observed between the VVI values and both comparators, the correlation was strongest between the VVI values and the image verified, known values (r = 0.863, p = 0.001). The validation results indicate that the VVI is a valid method which can be used as an improvement on standard viewshed analyses for the accurate representation of landscape visibility from a human perspective.