Development of an Autonomous Module for Weed Management in Agriculture
Abstract
In the state of Hidalgo, Mexico, agriculture faces critical challenges due to climate change, such as resource scarcity, while also striving to ensure food security. One of the most pressing issues in corn cultivation is inadequate weed management, which reduces productivity and increases production costs. This article proposes an autonomous system for weed detection and removal in corn fields, integrating artificial intelligence algorithms, real-time computer vision, and a robotic arm equipped with a laser and herbicide sprayer. Following a systematic review, the YOLOv8m model was selected for its real-time detection capabilities and its balance between accuracy and efficiency. The proposed system employs machine learning algorithms in Python, achieving over 80\% accuracy in distinguishing corn plants from weeds. Preliminary results demonstrate precise weed control, reduced environmental impact, and feasibility of field integration with minimal human intervention. This project marks a significant advancement toward more efficient and sustainable precision agriculture. As future work, the hybrid tool will apply selective treatments based on the size and location of the weeds, further minimizing herbicide use.
Keywords
Precision agriculture, machine learning algorithms, cornplant identification, weed detection