Photogrammetric captures (esp. orthomosaics) with object avoidance

[Don Engel]

A colleague suggested the ModalAI Starling 2 Max could be a good fit for my project and suggested I reach out here to see what the ModalAI engineers think of my use case.

I am hoping to use a drone to capture an orthomosaic of a cemetery with some trees and other obstacles.
The idea is to have a high-resolution aerial map, which I have done without trouble for cemeteries with short (or few) obstacles.

For a cemetery with tall trees, I would like to fly beneath their branches, while avoiding collisions with tree trunks, the graves themselves, etc.

Ideally, the "flight plan" would be a mission to get (nearly) full coverage of the target area, with decisions on where to re-visit (based on avoidances) happening mid-flight.

I expect this would require a long flight (swapping the battery regularly), lots of post-processing, and perhaps a live connection to a computer to help with the calculations.

I'm not sure if this is a novel thing to be doing or not (and if I'm going to write any code, I'd like to not be re-inventing a wheel), and would appreciate any guidance.