Agisoft Metashape

Agisoft Metashape => General => Topic started by: atedstone on November 01, 2018, 04:09:09 PM

Title: Generating time series of DEMs from a melting glacier surface
Post by: atedstone on November 01, 2018, 04:09:09 PM

I have a time series of UAV flight acquisitions (one flight per day for five days) from the Greenland Ice Sheet. This environment has a couple of key challenges:
1. The ice moves/flows at 0.5-1 m a day, so any GCPs laid out for the time series will be in a different geographic position at the end compared to at the start.
2. The ice melts, so GCPs laid on the surface will gradually lose elevation.

For point (1), my way around this (because I'm interested in comparing the same point on the ground at each time, even though it has moved in geographic space) is to leave the GCPs out for the entire time series, surveying them just once and using these same coordinates for all the flights - effectively cancelling out the ice motion.

However, point (2) is where I run into difficulties. Ideally I would let PhotoScan ignore the Z-coordinate of the GCPs because the Z-coordinate is only correct for a single time point and not the entire time series. I instead want Photoscan to estimate Z by itself, but still using my essentially fake X,Y coordinates.

In other words -  is it possible to disable GCP matching in the Z dimension?

(Note - there is no stable ground in these acquisitions on which to distribute GCPs - the nearest land is 30 km away!)

Title: Re: Generating time series of DEMs from a melting glacier surface
Post by: Dragline on November 02, 2018, 03:58:38 AM

That sounds like an interesting project!

I think this might lead you in the right direction... http://www.agisoft.com/forum/index.php?topic=7718.0 (http://www.agisoft.com/forum/index.php?topic=7718.0)

Another option might be running a drone with an RTK/PPK system, however your base station might be a long ways out. Try these guys: https://www.drobit.es/ (I believe he is member JMR here)

Or you could bring me and my RTK rig along...

Title: Re: Generating time series of DEMs from a melting glacier surface
Post by: atedstone on November 02, 2018, 05:08:03 PM

Hi Dragline, thanks for this. Useful post. I've tried its suggestions and unfortunately they don't really improve my Z solution, which suggests there is something more fundamentally wrong with inter-flight geo accuracy. Unfortunately we haven't any imminent chances to re-do the data acquisition but I will bear RTK/PPK in mind for the future.