Viewshed Analysis with Ridge Detection and Visualization on Digital Elevation Models

Master's thesis in Computer Science at NTNU

Written by:

• Nikola Dordevic
• Jørund Topp Løvlien

Supervisor:

• Theoharis Theoharis

Link to thesis

Abstract

In this thesis we examine the different viewshed analysis algorithms, namely R2, R3 and Reference planes. Using a ridge detection algorithm known as the Steepest ascent method, we extract ridge data from digital elevation models. The ridge data is then used together with the R2 and R3 algorithms to test whether this can provide speed ups. Along with this we implemented an rudimentary app to visualize the resulting viewsheds. In the end we found that using ridge data indeed does give speed up with the R2 algorithm. When it comes to the R3 algorithm with ridge data we found that our implementation proves to be much slower, which should be investigated further.

Results

Visualization of a viewshed analysis using the R3 Algorithm.

The resulting viewshed at 63.4162 degrees latitude, 10.3982 degrees longitude.

R2 Algorithm

R3 Algorithm

Reference Planes Algorithm

Installation

Install dependencies:
conda install -f environment.yml

Optional dependency installation for Anaconda environment:
conda install -c conda-forge geopandas
conda install -c conda-forge rasterio --freeze-installed

Controls for inspection tool in Panda3D:
https://docs.panda3d.org/1.10/python/debugging/inspection-tools/enhanced-mouse-navigation