An Improved Multiple Viewshed Algorithm for Region Partitioning for Filtering Candidate Viewpoints

Wang, Yiwen; Dou, Wanfeng

doi:10.1109/dcabes48411.2019.00026

Cited by 1 publication

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, a technique for reducing the number of viewpoints for the maximum coverage over an entire area (total viewshed), assuming that peaks are the ideal viewpoints for maximum coverage, has been developed [17]. Likewise, a similar approach toward applications where the total viewshed is the key issue continues to be developed [18]. However, there is limited research on systematic techniques, particularly those assessed with empirical data that identify how to select optimal key observation points, with a recent exception that addresses this shortfall [19].…”

Section: Introductionmentioning

confidence: 99%

Optimizing Viewpoint Selection for Route-Based Experiences: Assessing the Role of Viewpoints on Viewshed Accuracy

Openshaw

Chamberlain

2022

Land

View full text Add to dashboard Cite

A visual analysis is useful to assess potential impacts to our surroundings. There has been tremendous progress toward the optimization, accuracy, and techniques of these analyses. Viewshed analyses are a common type of visual analysis. The purpose of this study was to identify the optimal trade-off between the number of viewpoints needed to generate an accurate viewshed for a given route. In this study, we focused on identifying how a viewshed differs based on the sampling distance (interval) of viewpoints, topography, and distance of analysis. We employed the Geospatial Route Analysis and Visual Impact Assessment (GRAVIA) tool, a type of advanced viewshed that uses visual-magnitude measures. GRAVIA was applied across three different topographical environments (flat, hilly, and mountainous). We generated a one-mile-long segment for each environment and systematically discretized the route by varying the sampling-distance intervals from 1 m to 100 m. We also compared how the calculated results differed by distance from the route. The results showed a linear decrease in the correlation, though this was sensitive to the distance. When all distances were combined, a 30 m and 50 m sampling distance correlated to 0.9 and 0.7, respectively. However, when the results compared calculations beyond 300 m away from the route, the correlation values exceeded 97% for all the viewpoint-sampling distances. This suggests that for route-based analyses using visual magnitude, reducing the sampling rate can produce equivalent results with far less processing time while maintaining model precision.

show abstract

Section: Introductionmentioning

confidence: 99%