Methodology Development on Full-waveform Aerial LiDAR Data Analysis

Abstract

In this study, we proposed a robust full-waveform LiDAR data analysis technique which increases vegetation, canopy and ground detection compared to traditional discrete-return laser system. The full-waveform data is, at first, smoothed by a Finite-duration Impulse Response filter in order to remove high-frequency noise, and consequently, low-frequency noise is removed by setting a threshold. Peak detection is adapted to the improved full-waveform data by using Gaussian decomposition to detect reflective pulse returns. The quantitative evaluation of the increment in total returns and ground returns from the proposed method compared to the discrete-return laser system is performed. The variability in the ground point detection in different forest types and under varying seasonal conditions has also been analyzed. The influence of the increased ground point detection due to the full-waveform analysis in the accuracy of LiDAR derived Digital Terrain Model under varying seasonal conditions is also studied. Moreover, the multiple influences of seasonal conditions, forest types and topographical conditions are discussed in detailed.