PashuArogyam: Animal Disease Monitoring System Using Machine Learning

The frequent spread of infectious diseases among livestock and domestic animals remains a major challenge for farmers and veterinarians, especially in rural environments. Early detection is often difficult due to limited access to veterinary services and the reliance on manual inspection. This research presents a machine-learning-based disease monitoring system that uses image analysis and real-time prediction capabilities to identify visible symptoms in animals. A custom dataset was developed by manually capturing and labeling images using the CVAT tool. The detection model was trained using the YOLO framework, supported by preprocessing techniques such as normalization, augmentation, and resizing. The model integrates seamlessly with a Flask-based web platform connected to MongoDB for data storage. The system achieves high performance with 94.28% accuracy, 93.50% precision, and consistent mAP results. Experiments show that the system provides fast, accurate, and reliable predictions, making it suitable for field- level veterinary assistance and early disease awareness. This approach strengthens digital healthcare in the livestock sector and provides a scalable foundation for modern smart-farming technologies.