New tutorial : Aerial data processing with GCPs change

[dpitman]

The new tutorial (https://agisoft.freshdesk.com/support/solutions/articles/31000153696) completely omits the section on specifying high accuracy for the GCPs and low accuracy on the images.  And, unchecking the images before running Optimize Cameras.  This seems like a big change.  Is this a simple omission in the new tutorial or are those steps not recommended any longer?

I have a copy of the previous tutorial for comparison here: https://www.dropbox.com/s/k1yfx44z7zisoel/Metashape_GCP_workflow.pdf

[Alexey Pasumansky]

Hello Dave,

We will update the solution.

Actually nothing has changed, the accuracy for the reference coordinates or distances (cameras, markers, scale bars) should be defined according to the accuracy of the measurement device used.

And if the coordinate information is inaccurate or the coordinates should not be used for optimization/referencing for any other reasons, the corresponding elements should be unchecked in the Reference pane.

[dpitman]

Thanks for your response, Alexey !

That's good to know.

Regarding checking or unchecking the images, you said "And if the coordinate information is inaccurate or the coordinates should not be used for optimization/referencing for any other reasons, the corresponding elements should be unchecked in the Reference pane."

Does this mean that if the images are in a different coordinate system than the markers, you uncheck them.  But if they are in the same coordinate system, you leave them checked?

What I have been doing is importing the images (which are in WGS84, but then converting them to the state plane that the markers are using.  I then mark the images as less accurate and the markers as accurate (per the tutorial). 

Should I uncheck the images in this case or leave them checked?  So far I have unchecked them but maybe that is not necessary since I converted them to the state plane after they were ingested?  I'm just trying to understand what is happening under the hood when you uncheck the images vs leaving them checked even after you tell Metashape that the images are not accurate but the markers are accurate.

Thank you!
Dave

[Alexey Pasumansky]

Hello Dave,

In the recent versions of Metashape it is possible to select different coordinate systems for chunk, cameras and markers in the Reference pane settings dialog. But it is only possible when all the systems are defined on the same Datum or all of them have ToWGS84 parameters defined.

So it is not obligatory to uncheck cameras, for example, if their coordinates are in WGS84 and markers are in some UTM projection: you can just set UTM system for chunk and markers and select WGS84 for cameras. Effectively it would work similar to the case, when you convert the camera coordinates to the desired system in advance.

You can leave the cameras checked on in the pane, if the coordinate information for them corresponds to the selected coordinate system and the accuracy of the coordinate measurement is within the defined "Accuracy" value for cameras.
Note that you can set up individual accuracy for X, Y and Z coordinates, for example, if you are not confident in the altitude information from the drone, you can set 5/30, for example, which would mean that XY accuracy is 5 meters and altitude - 30 meters.

[dpitman]

Alexey, thanks so much for the details!

Do you know of anyone offering good in-depth instruction using Metashape for topographic work similar to what Simon Brown is offering for forensics?

[Alexey Pasumansky]

Hello Dave,

You can check for the Training Centers marked resellers on the corresponding page of our web-site:
https://www.agisoft.com/buy/resellers/

[dpitman]

Thanks Alexey.  Resellers in general usually have a general knowledge.  I was hoping for expertise in topography workflow using drones.

If anyone else knows of someone or YT channel, or whatever, please let me know.

Thanks,

Dave