So I’m in the midst of creating a new camera calibration for a 25mm perspective lens in Pix4d Mapper. My Easting and Northing are really good with an average of about 1cm. But my vertical bias is about 19cm. So the recorded position is 19cm higher than the computed position. And of course, my M2 helical antenna is about 19cm above the camera. Being this is an Arducopter and Pix4D setup, there’s no offset and lever correction. I could just adjust the vertical in global mapper, but with an accuracy of 2 or 3cm vertical for geolocation, my quality reports show up in red on the Z axis. So when someone looks at the reports, they think there’s something wrong lol. Here’s the crazy thing. I subtract the 19cm from my elevation in my geolocation metadata and I’m showing about 18cm lower than the calculated position! I realize there’s no lever adjustment accounted for, but I didn’t think it would completely reverse the error below the calculated Z. How does everyone else work around this that doesn’t fly DJI and uses Pix4D?
Can you please give some more details about your workflow? Do you record raw data logs on Reach M2 and then use them for processing? It’ll help me see if I can suggest any workarounds.
Yes, I’m using the raw UBX files for ppk in Emlid Studio. As for workflow, I’ve tried a number of things given Pix4D’s limited options of output geoid models. In small areas I would use 0m above ellipsoid in Pix image geolocation and the averaged above ellipsoid for my GCP’s. With bigger projects, I realized once my images are initially imported into Pix I could output the coordinates that Emlid studio gave me and send them to NRCan’s GPSH which converts all the ellipsoid elevations to one of our local Canadian models like CGVD28 HT2 or CGVD2013. Re-import the list and all my images are othometric. In that case, I just set the image and GCP/checkpoint vertical coordinates to arbitrary. Either way, if I select an image geolocation accuracy of anything under .5m, it shows up on the reports red as Pix4d is seeing the 18cm error of the antenna offset. I figure someone must have a workaround lol.
I’m not familiar with the Pix4D options. But I have a few ideas on how to handle it from Reach’s side. The first is to set the antenna height for the receiver in the Logging tab in Emlid Flow. You can do this if you’re recording raw data logs in RINEX format. This will add the antenna offset value to the raw data log so that the processing software can use it in calculations.
The second option is to set the antenna height directly in Emlid Studio. There is a field for this that appears after the log has been uploaded.
Have you tried this? Does it help?