Using GNSS antennas as GCP with tags/targets

PIX4Dmatic
1

I have an unconventional setup where I am installing antennas on the back deck of a ship and trying to calculate the relative X, Y & Z to a common reference point. These antennas are normally mounted at the ends of the container, so a consistent 20’ baseline

Typically we do not have access to RTK and instead log raw GNSS data and post process using PPK or PPP. Once processed out, we have the absolute X, Y & Z positions of the antenna and can then calculate the relative X, Y & Z to a common point.

We have dabbled with using PIX4Dcatch with a handheld device acquiring LIDAR/imagery, but whilst getting good results for vessel features the antennas themselves are rarely identified in the data. To that end, using the antenna’s themselves as ground control points has not worked.

Our next test was to be from above using a hired in guy with an Autel Evo II Pro drone or similar. Having found the downloadable catch tags we wondered if these tags can be used within PIX4Dmatic as well as PIX4Dcatch? If not we could just use a standard checkerboard design and mount this directly under the antenna (so minimal z offset).

Are there any issues with this methodology (bar the partial obscuring of the target with an antenna). For obvious reasons it is not logical for us to first define GCP’s using a separate GNSS system then install antennas at these locations, which once installed can be used to log data to generate accurate positioning for GCP’s…

Hope that all makes sense!

2

Hi, if the question was whether we could automatically detect the “Autotags” in PIX4Dmatic, like it’s currently done in PIX4Dcatch. Not yet for these targets, but this may be added, as we could run it with more computing power and hence find more matches. That said, this would work for sure for PIX4Dcatch data, but we have not done extensive tests yet on having those targets in drone images (a question of size of the target vs the image, etc…), but this would be nice, I agree.

That said, something that may help out in the meantime is the “AutoGCP” detection we have, with a classical checkerboard pattern, more here: Automatic target detection - AutoGCP algorithm

Hope that helps! Keep us posted on how it goes.

3

Thanks for that. Will make some over-sized checkerboard based on the GSD as per your link and test with those. I might also do some tags, but still on the fence if I am going to strap the iPhone to the drone at this point! Thanks again for the clarification and link to AutoGCP.

4

ha! it’s tempting, I suppose the drone would need to fly really slowly (and preferably have a gimbal), otherwise the images will be too blurry (yes, we tried to attach the iPhone to a few moving vehicles).