Could you clarify a bit about “inspection up tunnel”? I’m unsure if you’re wanting to put an ROV inside a fluid-filled tunnel and get a point cloud of the inside, or if you’re wanting to do a “scan from above” of the outside of a tunnel/pipe and possibly the ground around it.
Either way, the fundamentals of creating a point-cloud (or any kind of geometrically accurate map) requires a sensor that knows where it’s pointing, and a device that knows where it is, at least relative to where it was before. For an ROV to know where it is, for scanning the outside of a vertical pipe or pile it could be sufficient to have a depth sensor and accurate compass. For scanning a horizontal pipe you might need either an underwater GPS system, or a camera with computer vision processing that can determine relative position by aligning subsequent frames of a video feed together with the pitch, roll, and yaw of the ROV. For a very advanced sensor it can build up the map for you as it goes (as is done with several commercial LiDaR systems in air, notably in autonomous cars and some drones).
The requirements for your system depend quite a bit on the sensor you’re using, the required scan resolution, allowed scan time, and some additional factors. For the blueview in particular you might be interested in their webinar presentation on their 3D scanning capabilities, which might give you a better idea of what’s possible and perhaps also how to achieve it.
Definitely feel free to come back with more questions as you have them, particularly if you are struggling to understand some component of the process, or if you find a resource that you’d like clarification on
If you’re able/commercially allowed to, it’d be cool if you could document your progress in the Build and Dive Log categories of the forum