Tree Leaves Based Disease Prediction and Fertilizer Recommendation Using Deep Learning Algorithm

Abstract

The health of trees is a key component of ecological stability and diversity in ecosystems. Early detection of diseases that affect tree leaves can help with timely intervention and mitigation measures. The aim of this study is to determine whether or not tree leaves are healthy by evaluating high-resolution photos of the leaves. It offers an exclusive method for predicting tree diseases using deep learning—more especially, the VGG16 convolutional neural network architecture. The procedure entails gathering a substantial collection of images of tree leaves from various species and disease types. Improved robustness and generalisation of the model are achieved by applying data preparation techniques such as picture resizing, normalisation, and augmentation. Tree disease prediction is accomplished by customising the top layers of the pre-trained VGG16 model, which is used for feature extraction. To improve the performance of the proposed model, extensive training and validation processes are applied. The model's ability to classify illnesses is assessed using metrics such as accuracy, precision, recall, and F1 score. Developing a reliable and efficient tool to help environmentalists, foresters, and arborists quickly identify and address tree-related issues is the project's main goal. The study's findings provide an automated and scalable approach to early tree disease detection, advancing precision agriculture and environmental monitoring. The study supports sustainable practices for the preservation of global ecosystems by investigating potential real-world applications. Furthermore, extend the framework to provide information on fertilisers based on predicted disease.