We have built an Esri app to capture ground control points for our work, as it fits better with our workflows and ecosystem. The RTK corrections are done over Network RTK via NTRIP.
I’m getting some vertical deltas that I actually attribute more to it being in a remote area than an issue with the RS3, so I’m trying to apply corrections using the Stop & Go function in Emlid Studio.
I’m hitting a roadblock with the Project file (.csv) step: I can export my points form Esri as a .csv no problem, but when I then hit “Start” I get an error window.
Is there a .csv template I could use so I can make this work?
Hi @PointSetter, To process on the Emlid Studio, we suggest using the csv produced directly from the Emlid Flow. I have attached an example of the CSV. We can see that there should be a column with averaging time.
Do I understand correctly that you get the point in RTK Fix? This way, the stop & go post-processing should not fit into the workflow. Let’s investigate the height difference issue here. Could you share how many centimeters the difference is and how you compare it? Did you compare it with a known point?
Thanks!
1 Like
Yeah, I was worried about that.
Correct.
That’s what I thought, but with the deltas between points I wanted to try to correct them rather than having to go back out into the field.
It’s quite significant, approximately 0.80 meters (80 cm) in difference of the vertical values. I’m using the same GCS and geoid12b (also tried geoid18 and it was the same delta).
Did you record the raw data log and the RTCM3 correction? This way, we can then try to process it in Emlid Studio.
Could you share more about your workflow and which app you use? Did you use Field Maps?
If I understand correctly, the horizontal and vertical datum set on the app is align with the GCS?
I’m quite concerned, I just went out to a vertical benchmark, set the RS3 up and captured the point using BOTH the app I traditionally use (Esri Quick Capture) and Emlid Flow.
Once again, vertical values are off by 1.088 m!!!
Yes.
That’s what my original question was about!!! But because the points were collected from our Esri system and NOT Emlid Flow, I apparently can’t upload.
I’m attaching all the required information from today’s test, which still yielded a 1.088m error. The location is a NAVD88 geodetic control in Massachusetts.
I am also informing all teams to stop using our Emlid RS3’s until this is resolved. And not just this point corrected, but isolating the root cause of this issue.
Emlid Incorrect Vertical.zip (6.0 MB)
Additionally, when I tried to make the corrections using Studio, it said it couldn’t apply the changes to the .csv file that I exported from Flow.
Hi @PointSetter,
Sorry for not replying to your answer directly about stop and go since I was not so sure about the workflow before and want to have a better understanding of the workflow. I understand that accuracy is really crucial, and I am here to help to investigate this.
I reprocessed the data that you sent. I use the Reach base and Reach rover log. I see you also have CORS data attached. However, it did not cover the rover log time frame, so I didn’t use that.
I need to amend the averaging time of the point since in stop and go, we need to have an averaging time:
This way, the csv is able to be processed. In practice, I will try to avoid collecting points instantly and allow some time for averaging.
I’ve completed the postprocessing, but I’ll need an accurate base position point to compare the results. The RINEX header provides the base position from the latest epoch with meter-level accuracy. Could you share the exact base position?
Since it was obtaining CORS via NTRIP, I would go with the MAPL base station: NGS NCN Station Pages
@PointSetter, I reprocessed the data with the CORS data as a base and compared it with the NGS vertical control point you shared. I checked that there is a 25 cm difference vertically. I agree that this difference is too big and needs further investigation.
Is there a possibility there is another benchmark nearby to test?
Could you please collect some points, average them for 1-2 minutes for each point, and send the log for me to check?
Thank you.
There are more benchmarks, I’ll collect them in the near future (might not be until next week) and respond in this topic thread with the datasets.
Thank you. Please also record logs and RTCM3 data. I’d like to compare the RTK and PPK.
Sorry, other work has kept me in the office and I haven’t been able to go out to a site.
I’ll try to get the data collected tomorrow or Saturday, but if not it won’t be until late next week.
No worries. Feel free to let me know when you have a chance to test.
