Waveformlidar: An R Package for Waveform LiDAR Processing and Analysis

Tan Zhou,Sorin Popescu

doi:10.3390/rs11212552

Abstract

A wealth of Full Waveform (FW) LiDAR (Light Detection and Ranging) data are available to the public from different sources, which is poised to boost extensive applications of FW LiDAR data. However, we lack a handy and open source tool that can be used by potential users for processing and analyzing FW LiDAR data. To this end, we introduce waveformlidar, an R package dedicated to FW LiDAR processing, analysis and visualization as a solution to the constraint. Specifically, this package provides several commonly used waveform processing methods such as Gaussian, Adaptive Gaussian and Weibull decompositions and deconvolution approaches (Gold and Richard-Lucy (RL)) with users’ customized settings. In addition, we also developed functions to derive commonly used waveform metrics for characterizing vegetation structure. Moreover, a new way to directly visualize FW LiDAR data is developed by converting waveforms into points to form the Hyper Point Cloud (HPC), which can be easily adopted and subsequently analyzed with existing discrete-return LiDAR processing tools such as LAStools and FUSION. Basic explorations of the HPC such as 3D voxelization of the HPC and conversion from original waveforms to composite waveforms are also available in this package. All of these functions are developed based on small-footprint FW LiDAR data but they can be easily transplanted to the large footprint FW LiDAR data such as Geoscience Laser Altimeter System (GLAS) and Global Ecosystem Dynamics Investigation (GEDI) data analysis. It is anticipated that these functions will facilitate the widespread use of FW LiDAR and be beneficial for better estimating biomass and characterizing vegetation structure at various scales.

Highlights

The advent of Full Waveform (FW) LiDAR (Light Detection and Ranging) data including airborne and spaceborne have enabled new opportunities for vegetation structure characterization at a range of scales [1–5]
The present paper aims to introduce the waveformlidar package and its applications to R users
We incorporated commonly used FW processing algorithms with a new development of FW LiDAR data analysis, which is expected to alleviate the technical barrier of exploring FW LiDAR data and give users more flexibility to interpret results

Summary

Introduction

The advent of Full Waveform (FW) LiDAR (Light Detection and Ranging) data including airborne and spaceborne have enabled new opportunities for vegetation structure characterization at a range of scales [1–5]. Unlike Discrete-return (DR) LiDAR data which only store the crest of analog signal representing discrete echo, FW LiDAR data can store the entire echo scattered from illuminated objects with different temporal resolutions [1,6]. This advantage provides more information about the objects that the pulse interacts with and gives users more flexibility to interpret information inherent in waveforms [7]. Complicated processing steps and algorithms hinder the widespread use of Full Waveform (FW) LiDAR data To tackle these challenges, our package waveformlidar proposed several commonly used approaches and functions to conduct waveform processing and analysis such as Gaussian decomposition and deconvolution. With the aid of the package, these approaches can be implemented in the R platform and further relieve users’ concerns on complicated FW processing steps

Objectives

Methods

Findings

Conclusion