SilvaScenes: Tree Segmentation and Species Classification from Under-Canopy Images in Natural Forests

2025-10-10

AI Summary

Instance segmentation and fine-grained tree species classification in natural forest understories remain challenging due to severe occlusion, variable illumination, dense vegetation, and—critically—the absence of high-quality, domain-specific datasets. Method: To address this gap, we introduce SilvaScenes, the first large-scale, multi-bioclimate understory tree instance segmentation dataset, comprising 1,476 annotated trees across 24 species. Leveraging SilvaScenes, we jointly optimize instance segmentation and fine-grained classification using state-of-the-art deep learning architectures. Results: Our approach achieves an mAP of 67.65% for instance segmentation, validating the dataset’s utility; tree species classification attains an mAP of 35.69%, underscoring the inherent difficulty of fine-grained recognition in complex forest environments. SilvaScenes establishes the first benchmark and open-source resource for perception in forestry robotics, enabling systematic evaluation and advancement of understory-aware vision systems.

Technology Category

• Computer Vision: Scene Analysis & Understanding
• Machine Learning: Multi-instance/Multi-view Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Interest in robotics for forest management is growing, but perception in complex, natural environments remains a significant hurdle. Conditions such as heavy occlusion, variable lighting, and dense vegetation pose challenges to automated systems, which are essential for precision forestry, biodiversity monitoring, and the automation of forestry equipment. These tasks rely on advanced perceptual capabilities, such as detection and fine-grained species classification of individual trees. Yet, existing datasets are inadequate to develop such perception systems, as they often focus on urban settings or a limited number of species. To address this, we present SilvaScenes, a new dataset for instance segmentation of tree species from under-canopy images. Collected across five bioclimatic domains in Quebec, Canada, SilvaScenes features 1476 trees from 24 species with annotations from forestry experts. We demonstrate the relevance and challenging nature of our dataset by benchmarking modern deep learning approaches for instance segmentation. Our results show that, while tree segmentation is easy, with a top mean average precision (mAP) of 67.65%, species classification remains a significant challenge with an mAP of only 35.69%. Our dataset and source code will be available at https://github.com/norlab-ulaval/SilvaScenes.

Problem

• Addressing perception challenges in natural forest robotics
• Providing dataset for tree segmentation and species classification
• Overcoming limitations of existing datasets in forestry automation

Innovation

• SilvaScenes dataset for under-canopy tree segmentation
• Expert-annotated species classification across bioclimatic domains
• Benchmarked deep learning for instance segmentation performance