Chilli Leaves Disease Detection Using Thermal Images and Machine Learning Classification

The "Chilli Leaves Disease Detection using Thermal Images and Machine Learning Classification" project aims to address the critical issue of early disease detection in agriculture using advanced machine learning techniques. Leveraging a dataset comprising thermal images of chili leaves, this study explores the efficacy of DenseNet, ResNet, MobileNet, and VGG16 architectures for classifying four distinct leaf diseases: Bacterial Spot, Healthy Leaves, Leaf Curl, and Powdery Mildew. The dataset undergoes preprocessing, including resizing and normalization, to facilitate model training. Each model is trained using appropriate hyperparameters and optimization techniques to maximize classification accuracy. Additionally, a user-friendly web application is developed to enable farmers and agricultural stakeholders to input thermal images of chili leaves and receive real-time disease predictions. Evaluation metrics such as accuracy, precision, recall, and F1-score are employed to assess model performance. Results indicate promising capabilities in accurately identifying chili leaf diseases, with certain models exhibiting superior performance. This project contributes to the advancement of precision agriculture by providing a scalable solution for early disease detection, thereby facilitating timely intervention and improving crop yield and quality.

KEYWORDS: Chilli leaves, Disease detection, Thermal images, Machine learning, Classification, DenseNet, ResNet, MobileNet, VGG16, Bacterial Spot, Healthy Leaves, Leaf Curl, Powdery Mildew, Agriculture, Precision agriculture, Early detection, Crop yield.