Pb while processing big dataset Micasense Dual (10 bands)

I am contacting you because of processing issues with my Micasense Dual MX (using Trinity F90+ UAV).
When processing datasets, I have a kind of blury point cloud reconstruction leading to troubles while producing orthomosaic (“holes” at photo rigs locations, mismatching areas…) and time issues (> 2 months for processing a dataset of > 14000 images).
Example of “holes” in the resulting mosaic and report are attached.

Report link: FileSender
Dataset link (20.8Go): FileSender

I had an exchange with Quantum System support to check and avoid any problems with data quality (Data are correct, PPK processed, etc.). They failed in reconstructing the mosaic with Pix4D while they succeed with metashape in a couple of hours.
They mentioned a memory issue that does not allow to process data with a number of pics greater than 10000.
I already have processed datasets bigger than 10000 images (eBeeX + micasense MX - 5 bands).
Could the fact that processing 10 bands could be a limit ?

I really look forward any helpful comment…

Hi to those interested by this issue,
I had some feedbakcs from Quantum Support that is in touch with Pix4D support.
First explanation is the following:

Currently every of the 10 bands gets an individual geotag which is why Pix4D cannot process it since it is too many pictures. PixD will have a rig implemented that is only using the coordinates of one band and the then the other bands are aligned according to the offset in Pix4D.

They are exchanging to solve this issue…


Thanks for updating the community on your issue. While you wait for the update, I would like just to mention that you could try processing in https://www.pix4d.com/product/pix4dfields, our agricultural focused product.