Forum

Author Topic: PPP and shifting camera position or output position with RTK m300  (Read 1346 times)

Jake

  • Newbie
  • *
  • Posts: 6
    • View Profile
Hi,

We are flying the M300 with P1, LiDAR and Micasense multispectral sensors over forestry research trials.
We collect data on the same sites in different seasons, so need to get the best possible alignment.

We fly RTK flights with the above mentioned sensors, so get ~5cm relative reference. 
I would like to get the best absolute reference possible as well. 
We can PPP the base station logs https://webapp.csrs-scrs.nrcan-rncan.gc.ca/geod/tools-outils/ppp.php and get the absolute accuracy down to a theoretical 5cm as well.
We have also just begun using the Emlid Reach base and rover, so we would be using the Emlid studio wrkflow flow to shift the GCP locations.  We can also get updated camera locations from here as well.

I was just wondering if this is something anyone on here has a workflow for.

Do you shift each camera?  Or does shifting the output tif by the base shift produce the same results?

We are not interested in models, only output orthomosaics.

Thanks for any suggestions

Jake

Paulo

  • Hero Member
  • *****
  • Posts: 1338
    • View Profile
Re: PPP and shifting camera position or output position with RTK m300
« Reply #1 on: April 09, 2024, 02:40:49 PM »
Hi Jake,

In my book, I would use  PPP (minimum 4 h) to obtain a updated and precise base station coordinates,

Then calculate the shift from original to updated position and apply this to all camera reference coordinates and then redo the workflow to generate orto...

Hope this can be helpful,
Best Regards,
Paul Pelletier,
Surveyor

dpitman

  • Full Member
  • ***
  • Posts: 246
    • View Profile
Re: PPP and shifting camera position or output position with RTK m300
« Reply #2 on: April 09, 2024, 10:36:39 PM »
Hi Jake

Agree with Paul.

I would add that if you establish a good base position (via PPP) prior to using the base for RTK for your drone. The correction will be applied to the camera positions in real time.

If you are correcting your base after your drone flight, then you will need to also correct the camera locations. Emlid Studio can do that for you.

When you subsequently fly the same project in different seasons, if you can use your same corrected base position, then no corrections necessary.

The important thing is to get a good base position established.

Dave

Jake

  • Newbie
  • *
  • Posts: 6
    • View Profile
Re: PPP and shifting camera position or output position with RTK m300
« Reply #3 on: April 11, 2024, 01:13:18 AM »
Hi Paulo and Dave,

Thanks for the helpful replies.

Yes, I am logging for at least 5 hours before PPP.
Unfortunately we never have time to visit the site days ahead to survey, so we will be post processing the camera locations.

The Emlid workflow geotags the photos, but that wipes the yaw, pitch, roll.
 I assume that I will want this in order to best use Agisoft's photo alignment?
In that case I can pull the latitude and longitude from the Emlid Studio events.pos file into a csv.
The only problem here is there is no associated filename, I can write a little script that pulls the exif data from the original photos and then match them via timestamp.
I am worried this is getting a bit complicated and I missed a simpler solution.

I have two more questions..
1.  Can I do this when I have already built the ortho and update?  It doesn't seem likely, but just checking which step I need to go back to
2.  This seems to be a good solution for the  DJI P1 flights where I have the DJI .pos file, but we are also flying a Micasense MX dual.  There is no .pos file here, has anyone found a solution for this case?

Thanks very much again for the time

Jake

dpitman

  • Full Member
  • ***
  • Posts: 246
    • View Profile
Re: PPP and shifting camera position or output position with RTK m300
« Reply #4 on: April 11, 2024, 01:59:58 AM »

The Emlid workflow geotags the photos, but that wipes the yaw, pitch, roll.
 I assume that I will want this in order to best use Agisoft's photo alignment?

In that case I can pull the latitude and longitude from the Emlid Studio events.pos file into a csv.
The only problem here is there is no associated filename, I can write a little script that pulls the exif data from the original photos and then match them via timestamp.
I am worried this is getting a bit complicated and I missed a simpler solution.

Yes.  Metashape uses those parameters.  Emlid is aware of this and are actively working on Emlid Studio to retain all of that data.  Until then, you need to employ a workaround like you suggested.  You could also use RedToolbox (https://www.redcatch.at/REDtoolbox/) in the meantime. It has a nominal fee, and a fully functioning trial.


I have two more questions..
1.  Can I do this when I have already built the ortho and update?  It doesn't seem likely, but just checking which step I need to go back to

If you use the method to write the original exif, including the corrected camera positions, to the images. Then they will be used when ingested by MS.  If you use the original images, and then supply a camera position reference file, yes, you want to do all of that and get everything position wise sorted before having MS process any products.  MS can update the sparse cloud (tie points) based upon updating your changes.  But products beyond that need to be re-done.

dpitman

  • Full Member
  • ***
  • Posts: 246
    • View Profile
Re: PPP and shifting camera position or output position with RTK m300
« Reply #5 on: April 12, 2024, 10:40:44 PM »
Just a heads up. Emlid today just released Emlid Studio version 1.8 which retains the aforementioned metadata.
« Last Edit: April 16, 2024, 02:15:23 AM by dpitman »