Almost nothing in the built environment is as badly documented as what is under a street.
Records get lost, pipes get abandoned, connections get made and never drawn, and the only reliable way to know what is really down there is to go and look at it.
Going and looking is the whole problem.
A sewer is a confined space, and that is a legal category with a great deal attached to it.
Gas testing, rescue standby, harnesses, tripods, permits, a team above ground for every person below.
All of it before anybody measures anything.
Flying somewhere a satellite signal cannot reach
An ordinary drone cannot do this at all, and not because of the tight space.
It is because navigation underground has no external reference. No satellite fix, no landmarks, nothing to tell the machine where it is or how fast it is drifting.
So the aircraft has to build its own map as it goes, using a laser scanner to measure the walls around it and working out its own position from how that picture changes.
Do that continuously and you get two things at once: the machine knows where it is, and the map is the output.
The cage is the other half. A protective frame around the rotors means a knock against brickwork is a bump rather than a crash.
That changes how the thing can be flown, because the pilot no longer has to hold a perfect gap from every surface.
The result is a survey where the aircraft is allowed to touch things.
The pipe and the bridge above it
The sewer runs beneath a footbridge in the northeast of England that was due to be rebuilt.
Two questions needed answering before the design could be finished: what condition the pipe was in, and exactly where it ran relative to the new foundations.
A third question sat behind them. Surface water was getting in somewhere, either straight through the pipe wall or by way of a chamber, and nobody knew which.
The shaft down to it is 33 feet deep, and there are two of them.
The stretch of pipe between them measures 197 feet.
None of it is anywhere that a person would want to spend a working day.
Half a day against three
The work was done in the spring of 2025, by a survey firm working for the regional water utility.
The aircraft is a caged inspection drone from a Swiss maker, carrying a laser scanning payload, and it flew four missions: one to scout, one for close inspection, and two more for the mapping itself.
Two people and half a day covered it. The traditional approach would have been a full day of scanning followed by two days on foot inside the pipe.
Accuracy came out within 0.8 inches of ground based laser scans taken at the surface, which is comfortably inside what a design needs.
A gas sensor rode along and found nothing hazardous, which is worth knowing precisely because nobody had to go in to find out.
One of the directors put the gain plainly, saying they can now do four large manholes in a day rather than one.
What the comparison quietly leaves out
The time saving is real and the framing around it is a little generous.
Half a day covers the flying itself. It does not cover the processing afterward, which runs through several software packages before anything usable comes out.
The survey also did not replace the terrestrial scanning. It was checked against it, which is how anybody knows the accuracy figure at all.
And a drone sees surfaces. It does not sound a wall, take a sample or tell you what the brickwork is like behind the face.
For condition assessment that matters, in the way that a new method around a concrete repair or an autonomous truck tends to replace one step, as the maker’s own case study of the job records.
Why the water industry cares more than most
Sewers are the largest asset class that nobody can see, and the ones under old British towns are Victorian.
Every inspection that used to require entry required a team, a permit and a road closure, so most pipes simply never got inspected until something failed.
Lowering the cost of looking changes which pipes get looked at, and that is a bigger shift than the survey time on any single job.
The utilities pushing this hardest are the ones with the oldest networks and the most unrecorded connections, as the operator case study of the program describes.
The caged drone did not make the survey more accurate.
It made it cheap enough to do.
