Forum

Author Topic: Correlation matrix  (Read 6505 times)

kaarigar

  • Newbie
  • *
  • Posts: 5
    • View Profile
Correlation matrix
« on: April 06, 2018, 02:09:23 AM »
Is it possible to obtain or access the correlation matrix which is included in the processing report pdf? I scanned the python api document but I didn't see any api that exposes it. Thank you!

JMR

  • Hero Member
  • *****
  • Posts: 537
    • View Profile
Re: Correlation matrix
« Reply #1 on: April 07, 2018, 10:44:32 AM »
Yes, trhough camera calibration dialog, right click on camera and choose "distortion plot" you will find a tab with the correlation matrix.
Regards,
GEOBIT

kaarigar

  • Newbie
  • *
  • Posts: 5
    • View Profile
Re: Correlation matrix
« Reply #2 on: April 08, 2018, 08:15:37 AM »
Thanks for the reply, JMR. However I would like to access this correlation matrix information from the Python API. Is that possible? Thanks!

Alexey Pasumansky

  • Agisoft Technical Support
  • Hero Member
  • *****
  • Posts: 15321
    • View Profile
Re: Correlation matrix
« Reply #3 on: April 11, 2018, 02:17:57 PM »
Hello kaarigar,

Currently the correlation matrix cannot be accessed via Python API. We'll try to add the method to read it from script, but not sure, if the option would be included in 1.4.2 update.
Best regards,
Alexey Pasumansky,
Agisoft LLC

3DWinter

  • Full Member
  • ***
  • Posts: 103
    • View Profile
Re: Correlation matrix
« Reply #4 on: February 10, 2023, 02:30:20 AM »
Alexey,
Was the correlation matrix added to the python API since this last post?

What is the matrix below the parameters in
Code: [Select]
calibration.save()?
Thanks

Alexey Pasumansky

  • Agisoft Technical Support
  • Hero Member
  • *****
  • Posts: 15321
    • View Profile
Re: Correlation matrix
« Reply #5 on: February 10, 2023, 01:23:05 PM »
Hello 3DWinter,

You can access correlation matrix via sensor.calibration.covariance_params and sensor.calibration.covariance_matrix. The first one gives the labels of the corresponding columns and rows in the matrix.
Best regards,
Alexey Pasumansky,
Agisoft LLC