Infrastructure context
Potential targets include subsea cables, pipelines, foundations, moorings, harbour infrastructure, outfalls, seabed routes and environmental survey zones.
Inspection work often needs both visual context and positional accuracy. A useful system must know not only what it saw, but where it saw it, how reliable the capture was and whether the result is comparable with prior surveys.
AUV, ROV and USV roles
ROVs are typically remotely operated and often tethered. AUVs can execute autonomous underwater missions. USVs can support surface autonomy, communications, survey, relay or launch and recovery concepts.
These are not interchangeable categories. The right architecture depends on water depth, communications, endurance, sensor payload, risk tolerance and the required evidence standard.
Evidence required
Real subsea inspection claims require verified navigation, endurance, sensor calibration, procedures, operator competence and reporting accuracy. A concept page should not imply that those conditions already exist.
Keel's framing is that subsea inspection is a strong candidate domain for autonomous maritime intelligence because the environment is high-value, under-observed and operationally demanding.
Decision workflow
The useful output is an inspection record: imagery or sonar, location, timestamp, sensor provenance, condition notes, comparison with prior observations and a clear route to human review.
Keel treats the reporting and intelligence layer as part of the robotics problem, not an afterthought.
Operational caveat
Use cases describe potential applications for autonomous maritime systems and ocean intelligence. Operational claims require verified hardware, software, safety, regulatory and customer evidence.