Mission planning for heavily forested 80 acres on slope

Hey folks,

Asking for advice on mission planning.

Wanting to sort out rough orthos and 3d model of about 80 acres of forested property.

Known mission planning factors:

  • elevation of the terrain changes over this area, so if flown at one altitude, overlap on imagery will be variable
  • Lots of tall trees, so it might be hard to compose a nice 3D output

Been reading around including https://github.com/OpenDroneMap/WebODM/issues/6 and else where (DroneDeploy, Hivemapper, DroneHarmony, DroneMapper, etc) and I think I am settling on using Litchi as a mission planner and figuring out ODM for the processing.

Has anyone used Litchi for mission planning? Do they have reliable mission plan generator for 3D missions over varied terrain?

2 Likes

I have heard good things about Litchi, though I haven’t tried it myself.

For forest, especially over varied terrain, you will need lots of overlap, but also a fair amount of variation in flight angle. In the docs, we document how to, in a very robust way, ensure perfect camera calibration:

I need to rewrite this section, as there are easier ways to achieve good camera calibration. But this description of how to fly is very useful for forested landscapes, specifically performing a pair of cross-grids separated by 20°, ideally with a 5° forward facing camera.

Crossgrid overlap percentages can be lower than parallel flights. To get good 3D results, you will require 68% overlap and sidelap for an equivalent 83% overlap and sidelap.

As you have varied terrain, you may want each sub-flight (of the 8 subflights) to have 69 or 70% overlap / sidelap for overlap in excess of 83%

Doing so will give you really a really nice surface model and accurate orthophotos:

image




image

4 Likes

One thing I would add: if you use a flight planner that allows for terrain following, you can just fix your overlap and sidelap at 69 (83% equivalent).

2 Likes

Ok, great thanks for the input.

I did see the 5% of nadir shift in camera suggestion to act as a camera calibration of sorts but you also suggest it for complex terrain? And should every mission have the 5% shift or is it every second mission? And I was planning on doing two grids over the same area but you are suggesting eight? Or was it that each mission had eight sub-flights (eight sequences it was flying)?

1 Like

There’s no real penalty to always flying 5 degrees off nadir, and lots of benefits, though complex terrain is less subject to systematic lens errors.

Two grids, the second rotated 20 degrees from the first.

3 Likes

Ended up using DroneHarmony, after clipping a piece of the USGS DEM to utilize their terrain mission planning.

Screen Shot 2020-10-29 at 08.18.11

Hoping to collect this weekend and process the following week!

7 Likes

First attempt at an orthomosaic

Just ran:
docker run -ti --rm -v /Users/<myUserName>/datasets:/datasets opendronemap/odm --project-path /datasets project --dsm

Looks ok when zoomed out but there are some freaky seams and distortions in there :smiley: Exhausted, will have a deeper look into this tomorrow

Screen Shot 2020-11-01 at 20.32.37

2 Likes

Two things: increase your --texturing-nadir-weight to ~32, but also if you wait a couple weeks, you’ll likely see much better results. The 210 branch is :ok_hand:

4 Likes

By a couple weeks, I mean try it now. 2.1.0 is now in. Update everything and rerun.

Still set your texturing nadir weight high: somewhere between 28 and 32 should do the trick.

2 Likes

Not to revive a dead topic, but UGCS is a great tool for complex terrain mapping. Terrain awareness is excellent, and their support is top notch.

4 Likes

Thanks for the tip, looks cool!

1 Like

Did a big write up about my end-to-end journey with OpenDroneMap. Have a look if you have the time

6 Likes

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.