Showing posts from July, 2016

QGIS visibility analysis algorithm

It took me long time to develop an algorithm for the visibility analysis which would be comparable in terms of speed and quality to those available in other packages. And then some more time to tweak it for the type of analysis the QGIS visibility plugin is intended: higher volume calculation of multiple viewsheds over standard (coarse) elevation models. But here it is - an implementation which is faster than previous ones by a factor of 10 (at least).
1. Performance test There is a set of data files in plugin’s repository for testing purposes: a shapefile with several hundred observer points (462 to be precise) and an elevation model extracted from SRTM DEM, with 90 metres resolution (see resources below). All analyses were made using following parameters: a radius of 5000 metres for each point and eye level of 1.6 metres above ground. For ArcGis the standard viewshed module with multiple points was used (ref. below), and for QGIS plugin the cumulative visibility option…

QGIS viewshed plugin: a tutorial

I’ve had a remark recently that some kind of tutorial would be welcome for the visibility plugin. So here it goes…

The data which will be used can be downloaded from the plugin’s GitHub repo (link below). It comprises a DEM extracted from publicly available SRTM data (90 m resolution) and two sets of points (let’s call them A and B) which mostly correspond to archaeological sites I’m working on. The area in question is Istria (Croatia and Slovenia) and the projection is MGI Balkans 5 (EPSG : 31275).

A most basic use for the visibility analysis would be exploratory: would someone be able to see point B from point A? Such a query can be made by any viewshed algorithm available - but what about many observers from a number of points? In fact, when studying ancient landscapes we are often interested not only of what people could see, but also whether visibility influenced their preference for particular locations. For instance, is visibility a factor in the choice of settlement…

Advanced viewshed analysis: a Quantum GIS plug-in for the analysis of visual landscapes.

Published in: The Journal of Open Source Software 4(1). doi: 10.21105/joss.00032

Author : Zoran Čučković

Summary: Viewshed analysis is a standard feature of GIS software packages, such as ArcGIS (ESRI 2016), GRASS (Neteler et al. 2012) or ERDAS (2015). However, these implementations vary considerably in terms of their versatility and robustness. Software in the free domain is particularly poor in this respect: visibility analysis is generally implemented as a simple binary query (true/false) for elevation datasets (eg. GRASS or SAGA GIS). In order to meet the demands of a typical analysis concerning visual landscapes we would be interested to find out how deep are certain locations below the visible horizon, what is the overall visual potential of a landscape or which sites are connected in visual networks (cf. Higuchi (1983); Llobera (2003); Čučković (2015)).

Advanced viewshed analysis plug-in for open source Quantum GIS software has been made in order to meet some of these demands: b…