Canopy Structure Attributes Extraction from LiDAR Data Based on Tree Morphology and Crown Height Proportion

Urban green space is important for the well-being of urban residents. Seeking for three spatial dimension stereopsis is a very important issue in investigating urban green space. A potential applicability in the domain of urban tree space measurement and modelling has been explored based on LiDAR data in our study. This paper aims to present a framework—through a more automatic way—to extract canopy structure attributes. In this study, treetops were filtered by local maxima filtering algorithm from canopy height model. An improved spoke wheel algorithm was used to delineate the crown boundaries. And, an estimation issue of crown volume was simplified into three measurable parameters by estimating the crown structures. For accuracy assessment, data of 363 sampled trees located in the subset of Székesfehérvár city were selected randomly. The overall detection rate of treetop had proven to be 95.87% and crown boundaries were recognized effectively with a delineation quality of 88.59%, which were acceptable. About 80.26% of investigated crown volume estimates were obtained with shape distortion ranging from 3.1 to 7.8% according to the error analysis. The results indicated that the method can be used to extract canopy structure in urban areas.