I’ve been experimenting to find the perfect combination for generating orthophotos comparable to professional software. As you can see in the pictures above, I got decent results but there are some distortions and blurriness around vegetation and man-made objects(Point of interest are highlighted with translucent blue boxes). If anyone wants the orthophoto, here’s the link.
My System: Using latest docker nodeodm image on AWS ECS.
And here are the parameters that I’ve been working with:
"auto-boundary": true,
"cog": false,
"dem-resolution": 2,
"dsm": true,
"feature-quality": "ultra",
"force-gps": true,
"max-concurrency": 2,
"mesh-octree-depth": 12,
"min-num-features": 12000,
"orthophoto-resolution": 2,
"pc-geometric": true,
"pc-quality": "high",
"resize-to": 1600
If you also need the original images, do let me know. This orthophoto is the result of 36 images captured with DJI UAV.
I’ll be integrating lightning network in our project workflow, but I wanted to find out the perfect combinations that can cover a wide range of use cases, before we integrate the paid webodm service