Enhanced Crop Disease and Pest Detection Using Adaptive Neuro-Fuzzy Systems with ResNeXt Architecture

Le résumé fourni par la source

Crop pests and diseases are responsible for almost 40% of the loss in world agriculture; thus, their early and accurate detection is vital. An Adaptive Neuro-Fuzzy Inference System (ANFIS) with ResNeXt architecture is presented in this paper to improve crop disease and pest detection. ResNeXt explores deep hierarchical features, and ANFIS improves classification by adaptive learning. It has been tested on two benchmark datasets (PlantVillage and a$30,000+$-image realfield dataset) and it achieved 98.5% accuracy, 3-5% better than CNN-based methods. False detection rates were reduced by 12%. Its application to real-time precision agriculture with active disease management is suggested. Real-world deployment and computational speed are the future directions.

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