"Smart IoT Application for Rice Leaf Disease Classification with Optimized Deep Maxout Network"

What is it about?

Rice is the major crop in India. Early prevention and timely identification of plant leaf diseases are important for increasing production. Hence, an effective sunflower earthworm algorithm and student psychology based optimization (SEWA-SPBO) based deep maxout network is developed to classify different types of diseases in rice plant leaf. The SEWA is the combination of sunflower optimization (SFO) and earthworm algorithm (EWA). Initially, the network nodes simulated in the environment capture the plant leaf images and are routed to the sink node for disease classification. After receiving the plant images at the sink node, the image is preprocessed using a Gaussian filter. Next to preprocessing, segmentation using the black hole entropic fuzzy clustering (BHEFC) mechanism is performed. Then, data augmentation is applied to segmented image results and disease classification is done by a deep maxout network. The training of the deep maxout network is done using the proposed SEWA-SPBO algorithm. The proposed method detects the leaf disease more accurately with limited time and shows higher accuracy. Moreover, the proposed method attains higher performance with metrics, like accuracy, sensitivity, and specificity as 93.626%, 94.626%, and 90.431%, respectively

Featured Image

Why is it important?

The utilization of a combination of sunflower optimization (SFO) and earthworm algorithm (EWA) within a deep maxout network framework significantly improves the accuracy of plant disease detection. High accuracy, sensitivity, and specificity metrics indicate that the system can reliably identify various diseases, which is critical for taking timely and appropriate action, The ability to detect plant diseases early and accurately allows for prompt intervention, potentially preventing the spread of diseases and minimizing crop losses. This early prevention is vital for maintaining the stability of rice production in India, which has a direct impact on food security and the agricultural economy.

Perspectives

“

By enabling precise disease identification and classification, this method could reduce the need for broad-spectrum pesticide applications, leading to more sustainable agricultural practices with lesser environmental impacts. Accurate and timely disease detection can significantly reduce crop losses, directly benefiting farmers' incomes and reducing the economic burden of plant diseases.The success of this technology also hinges on its accessibility to farmers, including those in resource-limited settings, and the ease with which it can be integrated into existing agricultural practices.

”