Citation

Al-Badri, Ahmed Husham and Ahmed Salman, Ghalib and Mansor, Sarina and Arif, Arif Sameh (2025) Improved Detection of Urtica Dioica Weeds in Agricultural Fields Using YOLOv5 with Enhanced Non-Maximum Suppression. In: 2025 Multimedia University Engineering Conference, MECON 2025, 21 July 2025 - 23 July 2025, Cyberjaya, Malaysia.

Text 11.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

This study presents an enhanced deep learning approach for detecting Urtica weed plants using the YOLOv5 object detection model integrated with an Enhanced NonMaximum Suppression (ENMS) algorithm. The proposed ENMS+YOLOv5 model addresses key challenges in dense vegetation, particularly overlapping instances and finegrained morphological variations. A comprehensive evaluation was conducted using a dataset of Urtica weed plants across four growth stages—early, young, mature, and flowering—under varying environmental conditions. Performance metrics including True Positive Rate (TPR), False Negative Rate (FNR), Accuracy, and Intersection over Union (IoU) were used to benchmark the proposed model against baseline detectors such as DetectNet, AlexNet, SSD, and NMS+DLN. The ENMS+YOLOv5 model achieved an overall classification accuracy of 90.77%, the highest IoU of 91.21%, and the lowest FNR of 3.42%, demonstrating superior localization and detection performance. Visual analyses further confirmed the model’s robustness to occlusion, scale variation, and lighting conditions. These results suggest that the proposed method is highly effective for automated Urtica weed detection and holds significant potential for precision agriculture applications.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Deep learning
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Artificial Intelligence & Engineering (FAIE)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	18 Mar 2026 08:22
Last Modified:	19 Mar 2026 02:23
URII:	http://shdl.mmu.edu.my/id/eprint/15591

Downloads

Edit (login required)