Intelligent crop disease detection system using leaf image analysis with computer vision
- IEEE Xplore : 1-10
Résumé
Burkina Faso's agriculture sector faces major challenges, with annual crop losses reaching 40% due to plant diseases, affecting 2.7 million people in a situation of food insecurity. This research presents an innovative automatic plant disease detection system using computer vision, specifically developed for West African constraints. The system is based on the YOLOv11 (You Only Look Once) unified detection architecture, recognised for its optimal balance between speed and accuracy in real time, essential for mobile deployment to detect diseases in maize, tomatoes and chillies with an overall accuracy of ~99%. The image dataset from Kaggle has been validated by local agronomic expertise from INERA, ensuring the relevance of disease classes specific to the Sahelian context. The proposed architecture demonstrates superior performance to existing approaches while being optimised for mobile deployment. This solution contributes to the development of decision support tools for precision agriculture in West Africa
Mots-clés
Deep Learning, YOLOv11, Disease Classification, Precision Agriculture, Burkina Faso, Agronomic Validation