EXTRAÇÃO DE VARIÁVEIS DENDROMÉTRICAS DE FLORESTA URBANA UTILIZANDO DADOS DE LASER AEROTRANSPORTADO

Abstract

This study aims to evaluate the use of free and/or open-source tools to extract dendrometric variables from an urban forest using Airborne Laser Scanning (ALS) data. The study area corresponds to the Condomínio Parque Residencial Vivenda do Imbuí, located in Salvador, Bahia, Brazil, and was selected for its diverse tree population in an urban environment. The methodology consisted of three main stages: pre-processing of the LiDAR point cloud, extraction of dendrometric variables, and evaluation of result quality. Two methods were applied for tree detection: one based on the Canopy Height Model (CHM) and the other directly on the three-dimensional point cloud. To validate the results, a field survey was conducted using the Trees Count application, in addition to photointerpretation over an orthoimage. The variable extraction enabled the identification of the number, height, and canopy area of the trees, as well as the estimation of the vegetation cover in the condominium, which 5 totaled 31.83% of the study area. The results indicate that the CHM-based method showed greater adherence to the local reality, with a quality index (F1-Score) of 69%. The research highlights the potential of geotechnologies, especially LiDAR and the lidR package within the R programming environment, as efficient and replicable tools for managing urban forests, contributing to urban environmental planning and the promotion of quality of life in urban centers.