AI-Driven Sustainable Potato Disease Detection for Smart Agriculture: A CNN-WOA Hybrid Approach
Abstract
Plant leaf disease should also be detected and diagnosed early and correctly to improve the agricultural sector and ensure adequate food security and supply. The potato crops are extremely susceptible to various illnesses such as the early blight and the late blight and when they are not treated early enough, they cause significant loss of yield. Recent advances in the field of deep learning have demonstrated a tremendous possibility of automated plant disease diagnosis, which is the analysis of an image. The provided study suggests a relevant deep-learning-driven model of potato leaf disease classification using convolutional neural networks (CNNs). Plant Village is an online, publicly accessible dataset, which is used to train and evaluate this model and comprises of healthy, early blight, and late blight leaf pictures of potatoes. Pre-processing (resizing and normalization) also helps in producing additional feature extraction and model stability. It can be stated that the proposed CNN architecture can learn to discriminate the visual feature of a leaf image automatically and classify the images with a high rate of reliability (multi-class). The results of the experiment indicate that the offered solution is more accurate, precise, recalling better, and F1-score than the traditional machine learning and models of the deep learning that can be found in the literature. Based on the results, the CNN-based solutions are effective in smart agriculture and can be offered in real-time to assist farmers during the first stage of diagnosing the diseases and managing the crops.