Multi-Detection and Segmentation of Potato Seed on Conveyor Machines with Blur Conditions Using NAFNet and YOLO11

Nurhatinah Hr, Ingrid Nurtanio

2025

Abstract

In the agricultural industry, during the potato seed sorting process, automatic seed quality detection is a crucial requirement for enhancing production efficiency and consistency. This study proposes a potato seed detection and segmentation system, applicable to both sprouted and unsprouted seeds, by integrating the Nonlinear Activation Free Network (NAFNet) as the image pre-processing stage and YOLOv11s-Seg as the main detection model. NAFNet is used to reduce the blurring effect caused by conveyor movement, while YOLOv11s-Seg is employed to detect and segment potato seed objects. Experiments were conducted using video data at conveyor speeds of 0.70 m/s, 0.80 m/s, and 0.90 m/s. The evaluation results show that integrating NAFNet with YOLOv11 significantly improves performance compared to baseline YOLOv11s-Seg and YOLOv8. At 0.90 m/s, the proposed model achieved an mAP50 of 0.970, precision of 0.941, and recall of 0.890, outperforming YOLOv8, which only reached an mAP50 of 0.880 and a recall of 0.750. Consistent improvements were also observed at 0.70 m/s, where the system achieved an mAP50 of 0.988, precision of 0.968, and recall of 0.967. NAFNet effectively improves image quality and enhances YOLOv11s-Seg performance, offering substantial potential for accurate and reliable automation of potato seed sorting.

Paper Citation

in Harvard Style

Hr N. and Nurtanio I. (2025). Multi-Detection and Segmentation of Potato Seed on Conveyor Machines with Blur Conditions Using NAFNet and YOLO11. In Proceedings of the 1st International Conference on Research and Innovations in Information and Engineering Technology - Volume 1: RITECH; ISBN 978-989-758-784-9, SciTePress, pages 17-24. DOI: 10.5220/0014266700004928

in Bibtex Style

@conference{ritech25,
author={Nurhatinah Hr and Ingrid Nurtanio},
title={Multi-Detection and Segmentation of Potato Seed on Conveyor Machines with Blur Conditions Using NAFNet and YOLO11},
booktitle={Proceedings of the 1st International Conference on Research and Innovations in Information and Engineering Technology - Volume 1: RITECH},
year={2025},
pages={17-24},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0014266700004928},
isbn={978-989-758-784-9},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Innovations in Information and Engineering Technology - Volume 1: RITECH
TI - Multi-Detection and Segmentation of Potato Seed on Conveyor Machines with Blur Conditions Using NAFNet and YOLO11
SN - 978-989-758-784-9
AU - Hr N.
AU - Nurtanio I.
PY - 2025
SP - 17
EP - 24
DO - 10.5220/0014266700004928
PB - SciTePress