Deep Learning Approach for Mango Yield and Disease Prediction using Climate Data

Mrs. Beena K1, Manish Ray2, Manas3, Md. Hamza4

                                  Mrs. Beena K                                                                                                   Manish Ray
Department of Computer Science and Engineering,                   Department of Computer Science and Engineering,

 K.S. Institute of Technology (Affiliated to VTU, Belagavi). Visvesvaraya Technological University, Belagavi - 590018.                    Bengaluru, Karnataka - 560109, India.

                                     Manas                                                                                              Md. Hamza
Department of Computer Science and Engineering,                    Department of Computer Science and Engineering,

K.S. Institute of Technology (Affiliated to VTU, Belagavi).      Visvesvaraya Technological University, Belagavi - 590018. Bengaluru, Karnataka - 560109, India.

ABSTRACT- Mango cultivation in Karnataka faces significant economic challenges due to unpredictable climate variability and the prevalence of diseases like Anthracnose. Current agricultural practices often lack data-driven tools to accurately predict variety-specific yields or detect infections before they cause substantial losses. This paper proposes a unified decision-support framework that integrates Deep Learning and Machine Learning to provide a holistic solution for mango orchard management.The system utilizes a dual-model approach: a Convolutional Neural Network (CNN) with Transfer Learning for early-stage leaf disease detection and Random Forest or XGBoost algorithms for weather-based yield forecasting. Unlike existing platforms, this framework is specifically tailored to regional varieties such as Raspuri and Banganapalli, processing climate variables like rainfall and humidity alongside digital leaf images.