Unified Model for Crop Optimization: Leveraging Deep Learning and XGBoost for Optimized Crop Management

The agriculture industry requires innovations to enhance productivity, optimize resources, and mitigate risks from crop diseases and environmental factors. This research presents an AI-driven decision support system to assist farmers in crop yield improvement, resource management, and disease prevention. The system integrates multiple AI models for rice leaf disease detection, fertilizer recommendation, crop water level prediction, and crop selection. Data is collected from public sources, field experiments, and meteorological agencies. A Generative Adversarial Network (GAN) generates synthetic rice leaf images to enhance disease detection. Fertilizer recommendation is achieved using Feedforward Neural Networks and Random Forest, while Graph Neural Networks analyze soil moisture for irrigation planning. XGBoost is used for crop recommendation by evaluating environmental and economic factors. Deployed on a cloud-based platform with a Flask backend, the system enables real-time user interaction. It is evaluated using precision, recall, root mean squared error (RMSE), and classification accuracy. This AI-driven solution enhances agricultural decision-making, promotes sustainable farming, and supports food security.