![[2602.14140] Detection of On-Ground Chestnuts Using Artificial Intelligence Toward Automated Picking](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.14140] Detection of On-Ground Chestnuts Using Artificial Intelligence Toward Automated Picking
Summary
This article presents a study on using AI for detecting chestnuts on the ground to improve automated harvesting, highlighting the effectiveness of various object detection models.
Why It Matters
The research addresses the challenges faced by small producers in mechanized chestnut harvesting, offering a potential solution through advanced AI techniques. By improving detection accuracy in complex environments, this work could lead to more efficient and cost-effective harvesting methods, benefiting the agricultural sector.
Key Takeaways
- Traditional chestnut harvesting methods are costly and inefficient.
- The study evaluated 29 state-of-the-art object detectors for chestnut detection.
- YOLO models outperformed RT-DETR models in accuracy and inference speed.
- The best-performing model achieved a mAP@0.5 of 95.1%.
- The dataset and software from this study are publicly available for further research.
Computer Science > Computer Vision and Pattern Recognition arXiv:2602.14140 (cs) [Submitted on 15 Feb 2026] Title:Detection of On-Ground Chestnuts Using Artificial Intelligence Toward Automated Picking Authors:Kaixuan Fang, Yuzhen Lu, Xinyang Mu View a PDF of the paper titled Detection of On-Ground Chestnuts Using Artificial Intelligence Toward Automated Picking, by Kaixuan Fang and 2 other authors View PDF HTML (experimental) Abstract:Traditional mechanized chestnut harvesting is too costly for small producers, non-selective, and prone to damaging nuts. Accurate, reliable detection of chestnuts on the orchard floor is crucial for developing low-cost, vision-guided automated harvesting technology. However, developing a reliable chestnut detection system faces challenges in complex environments with shading, varying natural light conditions, and interference from weeds, fallen leaves, stones, and other foreign on-ground objects, which have remained unaddressed. This study collected 319 images of chestnuts on the orchard floor, containing 6524 annotated chestnuts. A comprehensive set of 29 state-of-the-art real-time object detectors, including 14 in the YOLO (v11-13) and 15 in the RT-DETR (v1-v4) families at varied model scales, was systematically evaluated through replicated modeling experiments for chestnut detection. Experimental results show that the YOLOv12m model achieves the best mAP@0.5 of 95.1% among all the evaluated models, while the RT-DETRv2-R101 was the most ac...