Forum

Author Topic: multi camera system approach for RGB + thermal sensor  (Read 2125 times)

Firenze

  • Newbie
  • *
  • Posts: 3
    • View Profile
multi camera system approach for RGB + thermal sensor
« on: July 06, 2022, 01:09:18 PM »
Hi all!

I'm trying to understand how multi-camera system and band layer works.

My goal is to obtain an RGB orthoimage and the corresponding radiometric orthoimage (the two sensors acquire "simultaneously") .

I organized the zoom sensor (jpeg) and thermal sensor (tiff) images into two subfolders and imported them as multi-camera system and set band layer 0 to the master rgb camera and band layer 1 to the slave thermal sensor

What I thought I could do was this:

-do the photogrammetric alignment with the RGB images only and build the orthophoto of the RGB images only;

-having done this, I thought that, since the sensors shoot simultaneously, with the same alignment and orthorectification parameters of the rgb photos, I could build the radiometric orthophoto without aligning the radiometric images between them (taking advantage of the photogrammetric alignment with rgb images, more precise).

having said that, I ran into two problems:

- the software tries to match rgb-rgb, rgb-tiff (even if the valid matches between rgb and tiff are zero), tiff-tiff; while, instead, I expected that, putting the images on two different layers, with rgb as master, the software would only align rgb-rgb and then used the same parameters to construct the radiometric orthophoto;

- consequently (I think) to the previous problem, the final result is: the rgb orthophoto is perfect, while the radiometric orthophoto is not (the frames are not merged well with each other and the result is bad).


I don't know if the approach is the right one or if what I thought I could do makes little sense


I hope this discussion can be interesting and that we can come up with some useful ideas.

Firenze

  • Newbie
  • *
  • Posts: 3
    • View Profile
Re: multi camera system approach for RGB + thermal sensor
« Reply #1 on: July 06, 2022, 07:41:50 PM »

I would like to specify that the rgb images are 5184X3888 while the thermal images are 640X512 and maybe this is real problem? I would like to know if there is a solution or not

Paulo

  • Hero Member
  • *****
  • Posts: 1303
    • View Profile
Re: multi camera system approach for RGB + thermal sensor
« Reply #2 on: July 08, 2022, 06:53:00 AM »
Ciao Firenze,

definitely that is the problem as the difference in resolution makes it imposible to link jpg with tiff...

best is to align in 2 chunks and transfer georeferenced markers from high resolution to radiometric to have common reference.

Best Regards,
Paul Pelletier,
Surveyor

Firenze

  • Newbie
  • *
  • Posts: 3
    • View Profile
Re: multi camera system approach for RGB + thermal sensor
« Reply #3 on: July 08, 2022, 09:33:08 AM »
thank you Paulo, to solve I did as you suggested. To conclude on the subject: if the images had had the same aspect ratio (even if with different resolutions), the process used would have been successful, or not?
« Last Edit: July 08, 2022, 09:34:40 AM by Firenze »

PolarNick

  • Jr. Member
  • **
  • Posts: 97
    • View Profile
Re: multi camera system approach for RGB + thermal sensor
« Reply #4 on: July 08, 2022, 12:21:27 PM »
You can't just replace images with images of another resolution (even of the same aspect ratio) because camera alignment includes camera internal parameters calibration and they are dependent on image resolution (and camera FOV, camera distortions and so on).

But I agree with Paulo that you can still reuse camera position and orientation (it will not be very accurate, but probably ok) if cameras were close to each other + were looking in the same direction + both cameras were taken photos at the same time. This script seems to do what you want - https://github.com/agisoft-llc/metashape-scripts/blob/master/src/transfer_orientation.py