Harnessing the Deep Learning techniques in Cashew plant disease management

Le résumé fourni par la source

The cashew (Anacardium occidentale K), originating in Brazil, was introduced to India by the Portuguese in the 16th Century. It is a significant crop for India, contributing substantially to the country's foreign exchange earnings. Various arthropods are known to infect cashew plants. This research employs pretrained deep learning and vision transformer model-based algorithms to identify pests and infections in cashew plants with high accuracy and precision at early stages. Cashew images from the CCMT (Cashew, Cassava, Maize, and Tomato) Dataset were used to train and develop advanced models. The dataset includes samples of Anthracnose, Gummosis, Leaf Miner, Red Rust, and healthy cashew plants/leaves. Vision Transformer model and pretrained DenseNet201 model were used to compare performance in detecting diseases in cashew plants. Hyperparameter tuning through random means improved model performance, achieving an accuracy of 91% and 95.95% for DenseNet201 and Vision Transformer models, respectively. Optimization using Differential Evolution on the Vision Transformer model further improved accuracy to 96.75%.

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