LEAF DISEASE DETECTION IN BANANA PLANTS FOR PRECISION AGRICULTURE

Abstract

Our lives are greatly impacted by the agricultural sector. The most signiftcant area of our economy is agriculture. Profttable agricultural goods are the outcome of effective supervision. Due to their ignorance about leaves illnesses, landowners harvest little. Since productivity determines proftt & loss, detecting plant leaf diseases is crucial. One of the main components of Indian agriculture is the production of bananas. At the same time, a prevalent issue in farming is that many diseases have affected the crop. Early disease detection is critical to crop management and banana output. Banana diseases cause losses that have a direct effect on the world’s fruit production and management sys- tem, which costs the nation money. To address these problems and help farmers avoid the disease in the ftrst place, a region-based separation using a suitable threshold approach and an adapted convolutional neural network are combined in the proposed method to enable banana disease detection and classiftcation. Recently, a CNN-free model on behalf of plant infection cataloguing has been used for computer vision tasks because it uses less resources during the train- ing phase and produces results that are identical to those of state-of-the-art CNN models. This hybrid model is established on a Transfer Learning-based prototypical monitored by a vision transformer (TLMViT). This study aims to present some deep learning approaches, such as support vector machines & con- volutional neural networks. Researchers may be inspired by this study to use the material to gain a deeper comprehension of associated disease prediction procedures.