Tree Detection and Diameter Estimation Based on Deep Learning

TL;DR

This paper introduces synthetic and real image datasets for vision-based tree detection and diameter estimation, demonstrating high accuracy of deep learning models and exploring their generalizability for autonomous forestry tasks.

Contribution

It provides new densely annotated datasets and evaluates deep learning models for tree detection and measurement, advancing autonomous forestry applications.

Findings

Deep neural networks achieved 90.4% precision in tree detection.

Models attained 87.2% accuracy in tree segmentation.

Results show promising generalizability and scalability of vision-based methods.

Abstract

Tree perception is an essential building block toward autonomous forestry operations. Current developments generally consider input data from lidar sensors to solve forest navigation, tree detection and diameter estimation problems. Whereas cameras paired with deep learning algorithms usually address species classification or forest anomaly detection. In either of these cases, data unavailability and forest diversity restrain deep learning developments for autonomous systems. So, we propose two densely annotated image datasets - 43k synthetic, 100 real - for bounding box, segmentation mask and keypoint detections to assess the potential of vision-based methods. Deep neural network models trained on our datasets achieve a precision of 90.4% for tree detection, 87.2% for tree segmentation, and centimeter accurate keypoint estimations. We measure our models' generalizability when testing it on other forest datasets, and their scalability with different dataset sizes and architectural improvements. Overall, the experimental results offer promising avenues toward autonomous tree felling operations and other applied forestry problems. The datasets and pre-trained models in this article are publicly available on \href{https://github.com/norlab-ulaval/PercepTreeV1}{GitHub} (https://github.com/norlab-ulaval/PercepTreeV1).