Deep learning-based detection of co-occurring diseases in mustard(Brassica juncea) crop using YOLOv11

Le résumé fourni par la source

Mustard [Brassica juncea (L.) Czern.] is an essential crop in agriculture, but its yield is often restricted by different kind of diseases, affecting both farmers and the edible oil industry. Recent advancements in artificial intelligence have transformed plant disease detection by enabling precise identification of various diseases. However, their detection and management become more challenging when crops are affected by multiple co-occurring diseases. To address this challenge, present study was carried out to implement deep learning algorithm for precise detection of co-occurring disease in mustard crop. A dataset consisting of 329 diseased leaf images of mustard crop with co-occurring symptoms of Alternaria blight and white rust diseases were collected during the 2023–2024 from the experimental fields of ICAR-Indian Agricultural Research Institute, New Delhi. Dataset processed through annotation and augmentation to boost detection accuracy of the network. A YOLO-based detection model (YOLOv11 and its variants) was trainedand validated. Their performance was evaluated using the test dataset. Among the YOLOv11 variants, YOLOv11-x showed the highest performance by achieving mAP @50 score of 96.2% with average F1-score of 93.8% on validation data, performing 4.8% relatively better than the least complex model (YOLOv11-n), highlighting its superior detection capability. Furthermore, YOLOv11-x also surpassed previous YOLO models (YOLOv8, YOLOv9, and YOLOv10) in detection accuracy. The experimental results demonstrated the effectiveness of YOLOv11 for co-occurring disease detection in mustard crops.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.