Canopy characteritzation using a LiDAR sensor

Abstract

This document explains the tools, algorithms and basically the whole process for the physical canopy characterization of a vineyard during three different stages of the vine growth. The main device used to scan the vegetation was a LiDAR LMS111-10100 and the programming language used to process the data is Python 3.6. The steps for this process described in this document are the protocol used in the field, the main idea of the first python code which prepare the LiDAR data to be readable for Cloud Compare (CC), how to clean the vines using CC, the basics of the algorithm used to compute the volumes of each vine, the useful coefficients obtained in this project and finally the results obtained. Apart from this LiDAR physical characterization, the vineyard was scanned using a drone which carries a multi spectral camera, Micasense Rededge. The frequencies registered are Blue, Green, Red, Red edge and Near-Infra-Red. A manually characterization was also done to compare the three methods This project aims to correlate the volumes extracted from the LiDAR data with the values given by the camera in the five frequency bands to see if it is possible to compute the vegetation volume from the photos, taking into account that the camera only has the top view of the vineyard. The Micasense RedEdge camera manually methods were already designed but the LiDAR method do not. Thus an implicit objective is to design and carry out a LiDAR protocol data. Regarding all the assembly, programming and results, it has been done successfully and efficiently. A direct and significant relation between the three main methods has been found. However, further research are needed in order to obtain an accurate method for canopy characterization using the proposed devices.