SIMLIDAR Simulation of LiDAR performance in artificially simulate orchards

Abstract

SIMLIDAR is an application developed in Cþþ that generates an artificial orchard using a Lindenmayer system. The application simulates the lateral interaction between the
artificial orchard and a laser scanner or LIDAR (Light Detection and Ranging). To best highlight the unique qualities of the LIDAR simulation, this work focuses on apple trees without leaves, i.e. the woody structure. The objective is to simulate a terrestrial laser sensor (LIDAR) when applied to different artificially created orchards and compare the simulated characteristics of trees with the parameters obtained with the LIDAR. The scanner is mounted on a virtual tractor and measures the distance between the origin of the laser beam and the nearby plant object. This measurement is taken with an angular scan in a plane which is perpendicular to the route of the virtual tractor.
SIMLIDAR determines the distance measured in a bi-dimensional matrix N M, where N is the number of angular scans and M is the number of steps in the tractor route. In
order to test the data and performance of SIMLIDAR, the simulation has been applied to 42 different artificial orchards. After previously defining and calculating two vegetative parameters (wood area and wood projected area) of the simulated trees, a good correlation (R2 ¼ 0.70e0.80) was found between these characteristics and the wood area
detected (impacted) by the laser beam. The designed software can be valuable in horticulture for estimating biomass and optimising the pesticide treatments that are
performed in winter.