Had a client reject our deliverable because they wanted sub-centimeter accuracy on a parking lot design survey. Our checkpoints showed 2-3cm which I thought was excellent for photogrammetry.
Is sub-centimeter even realistic with current drone technology? What techniques are people using to push accuracy below 1cm? More overlap? Lower altitude? Different processing settings?
Or is the client just unrealistic?
Sign in to join the discussion
Sign inSub-centimeter horizontal is achievable with good RTK, lots of GCPs, and low-altitude flights (150ft or less). Sub-centimeter vertical is much harder and generally not realistic for photogrammetry over large areas.
Your 2-3cm is actually very good. The client may not understand the limitations. I'd have a conversation about what accuracy they actually need vs what they think they want.
If they truly need sub-cm, they need LiDAR or terrestrial methods. Photogrammetry has physical limits. A well-executed drone survey with RTK and GCPs will give you 1-3cm. That's the sweet spot.
Anything claiming sub-cm from photogrammetry over a large area is either cherry-picking data or flying at absurdly low altitudes.
For parking lot design, 2-3cm is more than adequate. Most paving tolerances are +/- 1". Your data exceeds that.
Show the client the checkpoint report and explain what the numbers mean. A lot of times these accuracy complaints come from misunderstanding the data.