Dynamic viewshed analysis based on the digital elevation model: dynamic range proximity-direction-elevation reference line

Abstract

Viewshed analyses using Digital Elevation Models (DEMs) can provide decision support information in a variety of areas. Traditional viewshed analyses are mainly based on a circular visual range centred on an observation point, and lack the ability to dynamically adjust the visual range by view angle. In addition, these algorithms face challenges in achieving a balance between computational efficiency and accuracy. To solve the above problems, this study proposes a field-of-view (FOV) angular range based partitioning strategy to realise dynamic range proximity-direction-elevation reference lines (DPDERL), which improves the adaptability to complex terrains and the flexibility of viewpoint application environments. After partitioning, the algorithm is optimised for grid point search and reference line update. Experimental results show that the algorithm takes less than half the time of the Xdraw algorithm for a variety of terrain regions, has an error of no more than 5% compared to the R3 algorithm, and has less aggregation of error points than the PDERL and the XPDERL. The DPDERL approach provides an effective solution for large-scale dynamic range viewshed analysis by balancing speed and accuracy.