PRECISION AGRICULTURE THROUGH LEAF DIAGNOSTICS DISEASE DETECTION AND FERTILIZER RECOMMENDATION

Abstract

In order to maintain different ecosystems and for organic balance, the health of the tree is crucial. Early diagnosis of diseases affecting wooden leaves can facilitate timely intervention and molding efforts. This study presents a new method for predicting wood diseases using intensive learning, especially the architecture of a fixed nerve network. To determine whether wooden leaves are healthy or sick, this study analyzes images with high resolution of the leaves. A wide dataset that contains different images of wooden leaves represents different species and types of disease, gather as part of the process. To make the model more normal and stronger, data preparation techniques, including image size, normalization and growth. For functional extraction, we use a pre-trained CNN model called VGG16 algorithm. To work with the purpose of predicting our wood disease, we modify the top layers. To work in its best form of the proposed model, it is subject to extensive verification and training processes. The effect of the model in the disease classification is done using evaluation measures such as remembering, accuracy, accuracy and F1 score. A reliable and effective tool for environmentalists, foresters and arborists is the aim of the project to quickly detect and treat trees. By providing a scalable and automatic method for identifying early disease in trees, the findings from this study promote accurate agriculture and environmental monitoring. In addition, research delays practical applications that can help preserve ecosystems worldwide through permanent methods.