I'm hampered by having done aerial imagery acquisition and ortho-registration and -rectification in a bygone era. Designed and wrote a system using the Ashtech 3D GPS (4 receivers, 4 antennas in a diamond pattern atop aCessna 182, with a nadir-pointing video camera. The Ashtech provided data in RTCM and NMEA formats which were separately recorded, and the NMEA data were eventually used by audio-encoding on a VHS videotape system along with the images. We used 1-sec NMEA-0183 data bias-corrected for aircraft velocity over ground at a given timestamp, and used the 3D data to ortho-register the data.
I can't speak to the ublox solution save that, with one receiver you won't get 3D attitude info.
On Wed, May 20, 2020 at 3:03 PM Gary E. Miller <address@hidden
On Wed, 20 May 2020 15:08:32 +0200
"Felix Horvat" <address@hidden> wrote:
> thanks Gary for the quick changes. I wrote some comments on gitlab.
Replied to there. Please test the code.
> Is the bad data for fast readout also true for new multiband receiver
> like the ZED-F9P? They almost certainly don’t mention anything in the
> datasheet that would indicate lower accuracy besides a (much) higher
> power consumption.
Dunno. Someone needs to run real tests.
> The use case is kinematic PPP for aerial imagery on fast moving
The vague reports I have heard are for race cars doing 100+ mph.
What is your idea of fast? u-blox defines at least 10 categories
You may find this interesting:
The first paragraph states the obvious for anyone who's done high velocity kinematic work (no one in the survey world, at the time I did this stuff ever dreamed of 100 kt surveys, but we were able to accomplish it. Wish I'd published but it didn't seem novel when I was deep n the middle if it.
In general times, you want to remove ALL smoothing between fixes. In the GPS survey world, especially for static surveys, the norm used to be to decimate the data literally 10:1 to reduce autocorrelation. Code-phase position data is usually autocorrelated to smooth and provide a more aesthetically pleasing output, as the majority of humans don't want to see their results moving all over the place, which is a reflection of just how messy kinematic data really is.
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
address@hidden Tel:+1 541 382 8588
Veritas liberabit vos. -- Quid est veritas?
"If you can't measure it, you can't improve it." - Lord Kelvin