Multispectral Processing Changed in 2023.2.0

[RockRunEnterprises]

Just updated to 2023.2.0 and processed a multispectral mission from our Matrice 3M.

I've tried multiple attempts to re-process the project, but keep getting similar results with the seamlines present as pockets. This was an issue briefly in a previous version of D2M last year...

10-10-23

11-15-23

Same route, same settings, same everything.

I've noticed some other differences in the results from 2023.2.0 projects in orthos also...

[MarkBarker1]

Hi @RockRunEnterprises,

I had a look at the processing reports that you shared and did notice that your latest test is using the updated camera parameters that were added for the M3M drone. There are also a few slight differences in processing options between the two projects, but nothing that really stood out. We have some M3M data that I tested with the same settings and could not reproduce the same problem. Are you able to share out this dataset or another so I can test it on my end? Feel free to email me (mbarker@esri.com) or respond to this post and I can get a shared folder setup.

Thank you,

Mark

[RockRunEnterprises]

I just reprocessed the same dataset (11-15-23) on a different machine running 2023.1.1 with these results...

Notice the lack of prominent seam lines and the more uniform look.

@MarkBarker1 Feel free to set up a folder and I can transfer the images if you feel warranted.

To clarify on the original post, the dataset from 10-10-23 was processed on 2023.1.1 and the 11-15-23 dataset on 2023.2.0. I did try multiple attempts with various settings to clear up the seam lines and fractured look, but was unable to until I just processed again on the different machine.

[MarkBarker1]

Hi @RockRunEnterprises,

Thank you for that clarification. Yes, I believe at this point it would be good if I got a copy of the dataset on my end to test and see if I can produce the same behavior. Please let me know if you have any issues uploading the dataset. https://esriis-my.sharepoint.com/:f:/g/personal/mark8925_esri_com/EjwhKJaVqmhLpWLihjbqaHABXedIAtgfLO...

Regards,

Mark

[RockRunEnterprises]

Mark, I've uploaded the dataset. Not sure who marked the response as a solution...

[MarkBarker1]

Hi @RockRunEnterprises,

I ran through some testing with the dataset you shared and the main issue with the patchy results is due to the color balancing setting not being enabled in 2023.2 (Home tab> Options> 2D products> Color Balancing).

Can you try to enable it on your end with your test in 2023.2? 

Also, I noticed your screenshots are showing your ortho with what looks like a different band order (reddish instead of blue). Did you modify the band order or is that what you are receiving as a default output?

Thanks,
Mark