Mismatch between layers from the same camera - Pixel calibration?

[joebottoms]

Hello,

I'm new to these workflows but these forums have been very helpful in learning.

One thing I can't quite find the answer to, perhaps somebody can help:

I have been flying some rivers with a Zenmuse XT2 to produce four bands (RGB + FLIR) of imagery. However, when I export these for further analysis, I have a mismatch between the RGB and the FLIR layers (see attachment).

All the layers are processed to the same CRS (EPSG::3157 Canadian UTM Zone 10N), I had no NC errors as it appears all of the EXIF data is intact, but I am now seeing a mismatch in my layers.

I *think* this has something to do with my pixel size calibration, but I honestly don't really know.

I can't seem to find a good source on how to properly check that my particular camera is set up properly.

The camera specs for both the RGB and FLIR bands are attached, as well as the current auto-detected inputs from my camera calibration window.

Thanks for any advice!

[Paulo]

Hello Joe,

from the thermal camera specs, the pixel pitch is 17 um or 0.017 mm so i Would suggest you enter this value in your thermal camera calibration Pixel size in Metashape and do a Optimize.. see if it improves...

[joebottoms]

Hi Paul,

Thank you for the suggestion. I just optimized and re-processed the thermal layer using 0.017 for both the pixel height and width. Unfortunately, it didn't appear to fix the issue.

Would I perhaps have to re-process the RGB layers as well using those specs in order for them to match? Or is there something else I might be missing?

Joe

[Paulo]

Hello Joe,

nop your visual camera has a 1/1.7 inch format sensor with 12 Mpix, it should have something like 0.0019 mm pixel size...

[Paulo]

Joseph,
how are you prcessing the data set? As a camera rig with visual cam as master or ?

[joebottoms]

Forgive me, I'm not up on all the jargon in this field yet so I'll try here:

I flew single pass surveys with the XT2

I sorted the files into flir and rgb folders, and then imported them into metashape pro as a multi-camera system.

From there I simply followed the align > dense cloud > DEM > ortho workflow for each. The layers all seemed to stitch without any issue.

That said, I see that the RGB layer has unknown parameters for the pixel height and width and that the EXIF data wasn't actually found, so I was wrong to say it all successfully loaded in my earlier post. I didn't have any NC warnings though.

I don't know exactly what you mean when you ask if one is a master, but when I open the camera calibration window for RGB there is only one camera and when I open calibration for FLIR there are two, so I would guess that this is the case?