Fisheye? Does it work?

What fisheye lens/camera does WebODM work with?

I can get a acceptable result from my GoPro 10 what ever I do, I’ve tried with jpg converted from raw with really good detail.

I’m in need of help here.

I thought you had good results here: GoPro 10, first test!

I thought so but when I tried it outside, which should give even better result because of better light and with raw images giving better detail, it just doesn’t cut it.

The result is partial at best and there’s strange misplaced surfaces.

If it works inside and actually provides decent results and then gives partial results outside, then the problem is likely not the camera nor the processing approach but data collection problems. How are you choosing your photo number and ensuring you have enough overlap?

I take a photo and walk a few steps and take another one.

Maybe fisheye needs another approach in some way.

I vent back to basics!

Took photos around the living room again but in raw, converted to jpg and processed with both fisheye and auto, auto choose brown. Fisheye did best!

I used the resulting camera.json with the outside photos but it failed.

So now I’m a little confused.

How do I capture a part of a road in the right way with a fisheye?

I’ve read and seen people using a GoPro with success so it’s proven to work.

I just realised what I might have done wrong.

I just circled the road and took photos at the same hight, when I did the living room I took both low and high photos.

It’s clear I need to go back!

Human error was the issue!

I did a new test at a part of my own road and it worked well.


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