Dear all,
I’m trying to get a couple images from a Micasense MX (multispectral camera) orthorectified and aligned.
I tried both “fast ortho” and “default ortho” but in both cases the images are not aligned.
For example: the car is misaligned and there are lots of “white spots”.
Can someone help me? Is there a way to improve this alignment?
The files area available in this link: https://we.tl/t-9RmXrnrAD8
For others to see more easily:
I think part of the problem could be the lack of side lap? ODM perhaps should try to do a better job of handling linear datasets like this one. Have you tried to process it with some other software?
Also try to isolate a single band and process it individually. Do you get a better reconstruction in that case for a single band?
Hi @aperdigao - wellcome!
In pix4D processing the data as rig with all 5 channels, it seems to handle it better. Attached is the red edge.
Have you done any geo-tagging yourself or was all data already embedded in the original EXIF?
How does the 5 channel ortho look like in Pix4D?
Hi @gast,
Thank you for your input.
If I process each band i get quite good results.
My issue however is with the alignment of the bands.
Thanks,
Hi Piero,
Thank you for your input.
I think you might be right. I have lots of images but only processed a small sequence subset (that’s why they follow that path).
I will process the entire dataset and share the results.
Cheers
There are 5 individual orthos. All aligned. No single multi-channel ortho.
There were no errors that seemed to be related to the linear nature of the data set.
Well, I suppose that means we have work to do to improve things.
