Indoor view-based visibility analysis using UAV and TLS point clouds with Line-of-Sight correction

• Published in: International journal of applied earth observation and geoinformation 2024-05, Vol.129, p.103858, Article 103858
• Main Authors: Chen, Maolin, An, Aobo, Pan, Jianping, Mu, Fengyun
• Format: Article
• Language: English
• Subjects: Indoor perspective; Point cloud; Remote sensing; Viewshed image; Visibility analysis
• ISSN: 1569-8432; 1872-826X
• DOI: 10.1016/j.jag.2024.103858

•Indoor view-based visibility analysis method using point cloud is proposed.•Line-of-sight correction can address the influence of point cloud void on visibility.•Viewshed images reflect the real environment's impact on visibility.•Many other visual factors can be assessed by utilizing viewshed image. Visibility analysis is a crucial geographic information processing method, used to assess the observable spatial range from a particular location under specific conditions. It has broad applications, such as urban planning, landscape design and environmental research. Laser scanning provides detailed and accurate 3D data for visibility analysis, called point cloud. However, certain deficiencies still exist in current point cloud-based visibility analysis methods. Few methods are developed from indoor perspective and the influence of point cloud void is rarely considered. Firstly, indoor viewpoints are generated based on the windows extracted from terrestrial laser scanning (TLS) point cloud and unmanned aerial vehicle (UAV) point cloud is used to provide 3D representation of the environment around the viewpoint. Then, a half-sphere viewshed is created for each indoor viewpoint to analyze the visible spatial range. In this step, a novel line-of-sight correction method is proposed based on the semantic information from UAV point cloud, to correct the biased observation distance resulted from point cloud voids. Lastly, the volume of visible space is calculated based on the corrected viewshed image, to evaluate the visibility of each indoor viewpoint. Our method is validated on two datasets. The results indicate that our method enables more accurate simulation of visible space compared with previous studies by considering the effect of point cloud void. The proposed method can support the urban design from the perspective of indoor visual perception and accurate calculation of other visual factors in visual quality evaluation.