Mapping the 3D Distribution of Shorea Tree Species Based Upon Information Extracted from Worldview-2 and LiDAR Data

Abstract

Mapping and monitoring trees in tropical forest is deemed necessary especially for forest personnel. Unfortunately, acquisition of tree parameters for mapping purposes are tedious due to labour intensive, timely and cost-consuming. Thus, the advancement of remote sensing technology which provides tree parameters economically in term of cost and time saving over large forest area is in demand. The intent of this study is to map the distribution of Shorea tree species in the Ampang Forest Reserve using information extracted from Worldview-2 and LiDAR datasets. The pan-sharpening Worldview-2 imagery was used to classify the tropical trees using the support vector machine (SVM) image classification method. The overall classification accuracy for SVM method was 90.28% and the individual accuracy for Shorea and mixed tree species ranges from 68.25% to 82.86%. Finally, the classified result was overlaid with tree height information extracted from LiDAR data and forming the 3D distribution of Shorea tree species in the dense tropical forest area.