search

exeter-creww / SFM_Precision

A workflow to create precision maps in python
GNU General Public License v3.0
3 stars 1 forks source link

Confirm Marker accuracy settings with Mike #4

Closed h-a-graham closed 2 years ago

h-a-graham commented 4 years ago

In the documentation for Mike's script there is a suggetion that both tie point accuracy and marker accuracy should be set to the RMSE of tiepoint reprojection error. We have implemented this in the preprocessing function which will soon be integrated into the main script. Need to run this by Mike to make sure this is indeed correct.

piabenaud commented 4 years ago

If I can figure out how to edit the master branch... I will comment out 'chunk.marker_projection_accuracy = tiepoint_acc' as I don't think this is the correct interpretation of "appropriate values for ‘Image coordinates accuracy’ (these values should be the same as the RMS image error (in pixels) on the tie points and markers" but I'm not entirely sure if it should be the RMS of the marker in pix (it comes out on the photoscan report as image (pix) in the summary of control/check point rms) or if it should be what you reasonably think you can place it at i.e. 2 pix for UAV work or <0.1 pix for auto-detected circular markers and therefore just keep the values that you have used in your processing.

AndrewCunliffe commented 4 years ago

I agree with Pia, and have commented our this in the master script. We should still clarify this with Mike in the future.

h-a-graham commented 4 years ago

Okay - I've commented this out in the new module too. Shall we leave the issue open until we hear from Mike?

piabenaud commented 4 years ago

Yup