A Methodology for Forest Type Classification Using Aerial LiDAR Data

Abstract

In this study, we have developed a novel method for generating forest type map using airborne laser scanner data. An object-based approach is implemented for forest type classification, and feature images extracted from laser data are utilized for image segmentation and classification. Four types of feature images, namely, Digital Height Model, Reflectance Intensity, Ratio of First Pulse to Total Pulse, and Binary Reflectance Intensity are generated from the laser data. The first three ones are selected for image segmentation, and all the four feature images are used for classification. To assess the effectiveness of the proposed method, the classified map has been verified by comparing against a visual interpretation map. Our evaluation confirmed that by utilizing the proposed method we could achieve classification results close to the result of visual interpretation results.