workflow for marker based close range photogrammetry

[dejhost]

Hi.
I am relatively new to metashape, and I am trying to figure out a workflow for my historic close-range photogrammetry projects.

In my current sample project, I have 513 images, which cover a distance of about 70m. Each image shows a relatively high amount of coded targets, since I have deployed 1413 targets along the 70m.

Now I'd like to perform a bundle-adjustment based on these targets. I understand that it is not possible to fully exclude Tie Points, but following these instructions https://www.agisoft.com/forum/index.php?topic=14319.msg62976#msg62976, this is what I've done so far:

• Imported images
• Run alignment
• Detect coded markers
• Removed faulty markers manually (graphically in the model and using workspace by removing pointnumbers that weren't deployed)
• Run optimizeCameras
• set Marker accuracy to 0.2 pix and Tie point accuracy to 1+ pix

The attatchment show what a typical image and the model looks like at the current state.




My question is now, how to proceed?

• I have a set of 3D-points (real workld coordinates) that may be used for a camera calibration
• I have a list of known 3D-distances (also called "constraints" or "scale bars"), which I'd like to import. I know the manual way of creating scalebars in meatashape, but I wonder if I can just import the list, in order to speed up the process

Could please instruct me?
Thank you.

[Paulo]

Hello dejhost,

if you have known 3d points in your model, then definitely use them as GCPs in your optimization and it will improve the absolute georeference as well as camera calibration. How are your photos referenced (with GNSS or ....)?

As for scale bar use in API please look at https://www.agisoft.com/forum/index.php?topic=14526.msg63842#msg63842 as starting point...

[dejhost]

Thank you very much for your input, Paulo.

These 3D-points were generated in a lab, using a calibration wall and photogrammetry. Since this is an underwater project, no GNSS data available :-)

I imported the 3D-points as GCP's using File => Import => Reference
What would be the next step?
The markers that are no GCP's did not receive 3D-coordinates. That's at least what I see when I export the markers and look at the file.

When I choose "optimize cameras", it fails.
It seems I am missing a step in metashapes worflow.

---
I understand that I can import the scalebars one-by-one using the GUI, or through the python API. Cheers for this.

[Paulo]

Once imported into your project, you can select each GCP and with select Photos by markers context menu you can pin them in each photo where they appear and then Update transform and your project will be absolutely referenced according to these GCP coordinates...