Formation of a Lightweight, Deep Learning-Based Weed Detection System for a Commercial Autonomous Laser Weeding Robot
Problem
Weeds cause major crop yield losses (over PKR 65 billion per year in Pakistan), and autonomous laser weeding robots need accurate real-time weed detection to avoid damaging crops.
Why it matters
Reliable weed detection enables laser weeding robots to reduce yield losses and environmental risks of traditional weed management.
What the AI did
A lightweight YOLO5 single-shot object detector was trained on 9000 images from six farms in Pakistan covering three crops and four weed species.
Results
mAP of 0.88 @ IOU 0.5
Impact
Demonstrates dataset-level weed detection accuracy to assist laser weeding robots; field deployment not stated.
Record
| Venue | Applied Sciences |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Hafiza Sundus Fatima, Imtiaz Ul Hassan, Shehzad Hasan, Muhammad Khurram, Didier Stricker, Muhammad Zeshan Afzal |
| Citations | 41 (OpenAlex, at import time) |
| References | 39 |
| Model or method | YOLO5 |
| Organization | not stated in source |
| Primary source | https://doi.org/10.3390/app13063997 |
| Open access | gold |
| Keywords | weed detection, YOLO5, laser weeding, precision agriculture, object detection |
Importance rating
significant — thousands to millions affected
Abstract states weeds cause 11.5% yield loss worth over PKR 65 billion/year in Pakistan; robust detection enables laser weeding robots, reducing losses and environmental risks of traditional weed management.