Wild Berry image dataset collected in Finnish forests and peatlands using drones

2024-05-13

arXiv.org

AI Summary

To address the challenge of detecting wild berries—bilberry, cloudberry, lingonberry, and black currant—under highly variable illumination and severe occlusion in Nordic forest understories and peatlands, this work introduces WildBe, the first drone-acquired multispectral/RGB image dataset tailored for野外 berry detection. WildBe comprises 3,516 images with 18,468 PASCAL VOC–formatted bounding-box annotations across four berry species. It represents the first systematic effort to jointly acquire and meticulously annotate multi-class berries via drone-based remote sensing in complex natural environments. We benchmark six state-of-the-art object detectors (including YOLOv5, YOLOv8, and Faster R-CNN), achieving a mean Average Precision (mAP) of 32.7% under challenging forest-understory conditions. The dataset is publicly released on Hugging Face, filling a critical gap in Nordic wild-berry remote-sensing data and providing foundational support for sustainable, intelligent berry harvesting.

Technology Category

• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Berry picking has long-standing traditions in Finland, yet it is challenging and can potentially be dangerous. The integration of drones equipped with advanced imaging techniques represents a transformative leap forward, optimising harvests and promising sustainable practices. We propose WildBe, the first image dataset of wild berries captured in peatlands and under the canopy of Finnish forests using drones. Unlike previous and related datasets, WildBe includes new varieties of berries, such as bilberries, cloudberries, lingonberries, and crowberries, captured under severe light variations and in cluttered environments. WildBe features 3,516 images, including a total of 18,468 annotated bounding boxes. We carry out a comprehensive analysis of WildBe using six popular object detectors, assessing their effectiveness in berry detection across different forest regions and camera types. WildBe is publicly available on HuggingFace at https://huggingface.co/datasets/FBK-TeV/WildBe.

Problem

• Detecting wild berries in forests
• Using drones for berry harvesting
• Creating a dataset with diverse berries

Innovation

• Drones capture forest berry images
• Dataset includes diverse berry varieties
• Publicly available for object detection