The dataset contains files with sets of point coordinates derived via manual digitization of entire trees and all fruits on them. Twenty Bartlett pear and twenty cling-peach trees were digitized. Digitization was restricted to rigid, thick branches (> 2.5 cm) that present obstacles to machinery interacting with trees; robotic harvesting simulation and design were the targeted applications. Each branch segment that had small enough curvature was considered linear and was approximated by a conical frustum. Hence, an entire tree is represented geometrically by a set of connected frustums. The tree geometries and topologies were successfully encoded, and position RMS accuracy was better than 1.0 cm.

The fruit data was collected using Polhemus Digitization equipment in California commercial orchards, in the summer of 2017. The procedure of data collection and tree representation is the same as the one reported in:

Arikapudi, R., Vougioukas, S., Saracoglu, T. 2015. Orchard tree digitization for structural-geometrical modelling.
In Proceedings of the 10th European Conference on Precision Agriculture (ECPA), pp. 329 – 336, Volcani Center, Israel. https://doi.org/10.3920/978-90-8686-814-8

The Matlab files read the branch and fruit data from the Excel files and print the fruit trees for visualization. Each tree segment is represented as a conical frustum. More details are given in the Readme.txt file.