Résumé:
This master’s thesis focuses on weed detection using artificial intelligence (AI) in the field of agriculture. The aim of this study is to develop an
intelligent system capable of automatically detecting and classifying weeds
in agricultural fields through image analysis. Traditional manual detection
methods are time-consuming, costly, and prone to human errors. By leveraging advances in machine learning, image processing, and deep learning
algorithms, an AI-based system can provide accurate and real-time information on weed presence, enabling farmers to optimize their agricultural
production.The proposed system combines machine learning techniques
with image analysis and data processing to assist farmers in maintaining
crop health and reducing crop losses. This thesis presents the design, implementation, and evaluation of the system, demonstrating its effectiveness
in weed detection. The results show that the YOLOv9 model used for detection offers high precision and robustness in various conditions, although
some limitations have been identified. This study highlights the potential
of AI to transform weed management, promoting more sustainable and efficient agriculture.
