Kiwifruit tree modeling

Canopy reconstruction with LiDAR technology

This project was a collaboration with PlantTech Research Institute Limited, a company based in New Zealand. The goal was to create a software to measure canopy features of the kiwifruit orchards using a Point Cloud Data from LiDAR technology. The canopy features are the following:

​

• LIA: the Leaf Inclination Angles

• LAI: the Leaf Area Index

• LAD: the leaf area distribution

​

Such features are index of the wellbeing of the trees and can be used to create realistic simulations of kiwifruit orchards.

​

The algorithm was built with Python using Numpy, Pandas, Matplotlib, open3D, laspy and, Scipy as the core packages of this work. In order to find these features we had to classify points related to leafs, points related to the ground and, points related to everything else. To do this we perform various clustering approaches such as KMeans, DBSCAN and, Nearest Neighbors using the scikit learn package.

​

To test our results, we create a model of the kiwifruit tree with Blensor, a Blender software modification that allows to simulate LiDAR scanning.

​