Abstract
The availability of geospatial data is increasing daily, making the development of
applications that process data for large areas in high definition necessary. Viewshed analysis is
one of the applications that processes geospatial data to calculate the visibility of every point
of the digital elevation model from an observer for a predetermined area. It is an application
widely used in many scientific fields for projects with practical and/ or aesthetic value. For
example, analysing the disturbance of the field of view facilitates the recognition of areas,
whose visibility is not blocked by anthropogenic constructions. As a result, the detection of the
areas that exhibit rare natural beauty allows their exploitation for recreational purposes.
Regardless of the great benefits of the visual analysis, there is a major drawback in the usability
due to the requirements of advanced processors and the increased time required to derive the
results; thus, limiting their practicality. Aim of this study is to develop an algorithm that
optimizes high accuracy while minimizing the time necessary to achieve it.
The developed algorithm is based on Van Kreveld's [1] method, which is most
effective in reaching high performance alongside precision, when compared to similar
algorithms. Meanwhile, the interpolation technique is used in order to calculate the exact
altitude of every available spot, which leads to maximizing the precision of the algorithm.
Furthermore, we succeed in decreasing the execution time by implementing parallelization
techniques with the use of message passing interface (MPI). This results in an algorithm that
manages both precision and realistic execution time, which is practical for everyday use.
key words: Interpolation technique, Van Kreveld’s algorithm, Viewshed analysis,
Parallelization, Message Passing Interface, Line of sight