Forum

Author Topic: PPK or RTK based UAV  (Read 1186 times)

gunen

  • Newbie
  • *
  • Posts: 2
    • View Profile
PPK or RTK based UAV
« on: November 20, 2017, 12:57:57 PM »
Hi,
I use PPK based UAV. Actual image coordinate are written in EXIF. I have obtained the actual coordinates of the image by using the PPK method and want to produce the Point Cloud without using GCPs. Which paths should I follow different from marking GCPs? Should I change the camera accuracy (m) and camera accuracy (deg)? Is there a suggested value of accuracy? Which steps should I follow beside the classic process to produce a precise Point Cloud without the ground control point? Thanks for now.

SAV

  • Hero Member
  • *****
  • Posts: 651
    • View Profile
Re: PPK or RTK based UAV
« Reply #1 on: November 21, 2017, 06:40:29 AM »
Hi gunen,

Yes, you will need to change the accuracy settings to match the accuracy of your PPK system. I guess a value between 0.02m (= 2cm) and 0.15m (= 15cm) for camera accuracy is realistic using PPK (will depend on the 'quality' of your system though). Leave the camera accuracy (deg) at the default value.

Make sure that the cameras are checked in the REFERENCE PANE to guarantee that they are used as a reference.

After image alignment you should always 'clean' your sparse point cloud using EDIT > GRADUAL SELECTION. Remove all points with high reprojection error (choose a value below 1, I suggest to use 0.5-0.8 ) and high reconstruction uncertainty (try to find the 'natural threshold' by moving the slider). After you have removed 'rogue' tie points from the sparse point cloud, run the optimize camera alignment step.

To verify your model's accuracy, you could/should include a few scale bars. Simply use a tape measure to get the distance between obvious points/objects before you fly. You could put down a few GCP markers and simple measure the distance between them, for example. In Photoscan, add them as check scale bars (not control scale bars), which means they are NOT used to optimise your tie points (sparse point cloud)  and camera locations and orientations. Hope that makes sense.

All the best.

Regards,
SAV