Autonomous Aquaculture Net Inspection and Patching on a Low-Cost ROV

Abstract

Undetected holes in aquaculture net cages can cause massive fish escapes; however, autonomous inspection and repair of these structures remains an open problem. Existing systems usually focus on inspection only or require teleoperation for patch deployment, leaving the detection-to-repair cycle only partially autonomous. This paper presents an autonomous pipeline implemented on a low-cost, general-purpose ROV, the BlueROV2 Heavy, capable of closing this cycle in a controlled environment without operator intervention after mission start. The vehicle visually searches for the net, estimates its relative position using onboard perception, inspects the surface, detects holes that satisfy a predefined size criterion, and executes the patching maneuver. The system was developed and tested in the Stonefish simulator and experimentally validated end to end in the indoor CIRTESU test tank. The main contributions are end-to-end autonomy on an affordable robotic platform and terminal patch attachment through closed-loop position tracking, improving deployment repeatability under moderate underwater disturbances.