Urban tree mapping with aerial lidar

Abstract

Recent advances in remote sensing technologies enable rapid estimation of some physical characteristics of the vegetation, as well as to distinguish species of urban trees and analyze spatiotemporal distribution, which results in significant improvements in mapping accuracy. By using aerial LIDAR data, height and crown diameter of individual trees were estimated in an urban area south of Monterrey, Mexico in order to design an efficient mapping methodology for urban trees. 1) semi-automathized design filters for extracting points associated vegetation and 2) use of contour lines on a digital surface model, two extraction techniques were used. The results were verified with field data obtained during the same period belonging to the Arboretum project Tecnológico de Estudios Superiores de Monterrey, and obtained with traditional methods of forest measuring 57.80 hectares with 2 575 trees reported. Canopy mapping was efficient for the 74.7 % of individuals despite registering a 6.4 % measurement error in geoposition and inventory measurements according to the comparison of the position and crown width LIDAR data versus 2010 inventory. Results showed similar efficacy between field sampling at the point cloud and aerial LIDAR (x = 0.70 m, z = 0.10 m) according to linear regression analysis of height (R2= 0.885) and crown width (R2= 0.902). The best estimates were for Quercus shumardii (95 %) and Quercus vaseyana (94 %).