A robotic vision system to measure tree traits

TL;DR

This paper introduces RoTSE, a robotic vision system that accurately measures tree traits like branch structure and angles in field conditions, aiding robotic pruning and phenotyping.

Contribution

The paper presents a novel robotic vision system for in-field measurement of tree traits, integrating image acquisition, segmentation, reconstruction, and trait computation.

Findings

Accurate measurement of branch diameter with RMSE of 2.97 mm.

Branch length estimation with RMSE of 136.92 mm.

Average processing time of 8.47 minutes per tree.

Abstract

The autonomous measurement of tree traits, such as branching structure, branch diameters, branch lengths, and branch angles, is required for tasks such as robotic pruning of trees as well as structural phenotyping. We propose a robotic vision system called the Robotic System for Tree Shape Estimation (RoTSE) to determine tree traits in field settings. The process is composed of the following stages: image acquisition with a mobile robot unit, segmentation, reconstruction, curve skeletonization, conversion to a graph representation, and then computation of traits. Quantitative and qualitative results on apple trees are shown in terms of accuracy, computation time, and robustness. Compared to ground truth measurements, the RoTSE produced the following estimates: branch diameter (root mean-squared error $2.97$ mm), branch length (root mean-squared error $136.92$ mm), and branch angle (mean-squared error $31.07$ degrees). The average run time was $8.47$ minutes when the voxel resolution was $3$ mm$^3$.