Smart Mango - Smart System for Mango Plantation

This research paper introduces an innovative approach utilizing machine learning to revolutionize mango cultivation practices. The study aims to enhance productivity, quality, and disease management in mango farming through four main objectives: creating an automated mango grading system based on weight and color using machine learning techniques, accurately predicting mango yield using historical data and weather patterns, forecasting optimal water requirements for different growth stages of mango trees, and implementing an early disease detection system. The research utilizes transfer learning with VGG16 and ResNet50 models for image-based tasks, SVM for grade prediction, and linear regression for yield and water level prediction. These machine learning models help in achieving promising results in grading accuracy (88 percent) and disease prediction, though further data exploration is needed. Yield prediction accuracy reaches up to 74.9 percent, and water level prediction attains percent, underscoring the potential of machine learning to revolutionize mango cultivation for more efficient and sustainable practices.