Tree Species Identification from Bark Images Using Convolutional Neural Networks

TL;DR

This paper introduces BarkNet 1.0, a large dataset of bark images, and demonstrates high accuracy in tree species identification using convolutional neural networks, advancing forestry applications.

Contribution

The creation of BarkNet 1.0 dataset and the application of deep learning for accurate tree species recognition from bark images.

Findings

Achieved 93.88% accuracy on single crop

Achieved 97.81% accuracy with majority voting

More training data from diverse trees improves recognition

Abstract

Tree species identification using bark images is a challenging problem that could prove useful for many forestry related tasks. However, while the recent progress in deep learning showed impressive results on standard vision problems, a lack of datasets prevented its use on tree bark species classification. In this work, we present, and make publicly available, a novel dataset called BarkNet 1.0 containing more than 23,000 high-resolution bark images from 23 different tree species over a wide range of tree diameters. With it, we demonstrate the feasibility of species recognition through bark images, using deep learning. More specifically, we obtain an accuracy of 93.88% on single crop, and an accuracy of 97.81% using a majority voting approach on all of the images of a tree. We also empirically demonstrate that, for a fixed number of images, it is better to maximize the number of tree individuals in the training database, thus directing future data collection efforts.