Assessing the visibility of urban greenery using MLS LiDAR data
Liyu Tang,Jianguo He,Wei Peng,Hongyu Huang,Chongcheng Chen,Can Yu
DOI: https://doi.org/10.1016/j.landurbplan.2022.104662
IF: 8.119
2022-12-22
Landscape and Urban Planning
Abstract:Given the ecological, cultural and psychological functions of urban vegetation, a growing number of studies have focused on daily accessible greenery visibility. Mobile laser scanning (MLS) can rapidly obtain dense point clouds that can be used to extract vegetation. To address this issue, we propose a novel method for calculating the green view index (GVI) based on MLS three-dimensional (3D) point clouds. In our study, the GVI was specified as the ratio of greenery viewing angles to the total number of viewing angles in view. The GVI calculation procedure was as follows. First, the vegetation points were extracted using the density-based spatial clustering of applications with noise (DBSCAN) algorithm and the PointNet++ deep learning algorithm. Second, based on the GVI specification, a virtual camera was constructed in a 3D point scenario to estimate greenery viewing angles in view and generate depth images, and then, the GVI value was calculated. This method is flexible and can be used to calculate the GVI at any site with any direction where 3D point scene data are available, and thus, it is suitable for evaluating various types of urban greenery. We conducted a case human-centered assessment of road greenery in a partial area of Jinshan District in Fuzhou, China, based on MLS point clouds, evaluating visible greenery and analyzing the relations among the GVI, greening pattern, and road green belt mode. The results showed that the overall visible greenery in the study area was good and that the GVI value of most road sections was more than 15 %. The method has potential for urban green space planning and management.
environmental studies,ecology,geography, physical,regional & urban planning,urban studies