AI-Driven Fish Health Monitoring and Recommendation System for Aquaculture

Abstract

Aquaculture is essential for global food security but is highly affected by disease outbreaks. This paper presents FishCare AI, a real-time aquatic health monitoring system using deep learning and conversational AI. The system employs a dual-stage YOLOv8 architecture, where the first stage detects fish and identifies species, and the second stage performs disease detection only when required, improving efficiency and reducing latency. A context-aware chatbot provides reliable, evidence-based recommendations. The system achieves 92.4% accuracy in species detection and 88.7% in disease detection, offering a scalable and efficient solution for practical aquaculture monitoring.