This research outlines the integration of image processing with an Aquabot, underwater rover to monitor the growth of corals. Aquabot, a tethered underwater rover with a floating communication/power station, a Python/C# ground-station dashboard, and a cloud back end for real-time telemetry and image analysis focused on coral monitoring around Sri Lanka. The platform integrates dual-camera video, pH/temperature/depth sensing, joystick teleoperation, and an inference pipeline for coral varieties and coral disease detection trained on locally curated, polygon-annotated datasets. On held-out tests, the varieties detector achieved precision = 0.679, recall = 0.558,  = 0.596 and :0.95 = 0.425 (124 images/492 instances), while the disease detector achieved precision = 0.346, recall = 0.331,  = 0.295 and :0.95 = 0.188 (121 images/489 instances). Field evaluations demonstrated stable maneuvering, synchronized data capture, and ~ 12 min endurance at peak load, validating the end-to-end pipeline from acquisition to analysis. The system provides a practical, extensible baseline for sustained coral monitoring and conservation workflows in shallow coastal environments.