Help with GCPs here

Reach RS2 / RS2+ RTK / PPK configuration
1

Hi guys,
i’m a newbie here.
I usually fly my drones for visual inspections only but lately I was able to test the RS2 I was able to buy used for a fair price and using GCPs in some test missions.

    • Well, as far as I could understand from reading around:
      in order to have good results, I should operate with a (1) base and (2) a rover.
      BUT… the Base can’t be positioned anywhere, but on special location which have been already mapped before and are “known” points (OR… by using long time PPP data on them with Rinex logs)
    • There are several “datum” for several countries / regions and it is especially important to study ellipsoid and geoid differences and NOT only WGS84 that’s usually seen around. For Italy for instance, WGS84 must not be used for surveys. Height is also another datum included.
    • Not having 2x Emlids, I should connect and collect my GCP by using NTRIP connection RTK and use my Emlid as a “rover”. The NTRIP RTK provider will create a virtual base (VRS 30) which will be calculated by this NTRIP connection to.
    • I subscribed to an GNSS NTRIP RTK service here, which I have now 30 days free before buying this service here in Italy…so I have a few days to test before spending money on this. They use RTCM method.
    • I was able to map 5 GCPs in a mission… I put in my EMLID as a “Reference System” - WGS84 UTM 33N… and NOT using the local datum (ETRF2000) used here in Italy which is used by this NTRIP provider as written in their home-page (as I couldn’t find in the list?)
    • I was able to export to .csv file the results and I can “see” the WGS84 “N” and “E” locations of the 5 GCPs I took.
    • If I put this “log” into “WebODM-GCP setup” (like Pix4D, but free) I can actually see the 5 points clearly and they are recognized!
  8. Then I, for each GCPs, manually click every single GCP I collected and associate it to the “photos” I took (as explained in the WebODM manual)
    • I then elaborate the map and… BOOM… the 2d orthophoto is absolutely wrong and stretched for miles (!)… like the GCPs were not used or wrong 100%…

QUESTIONS

    • I’m at a loss here… what is wrong in my procedure? Does it make any sense?
    • Which setting should I put in my EMLID Reach known that my NTRIP provider writes that they use ETRF2000 Roma40 “Reference system”?
  3. Should I select in the EMLID -----) WGS84 UTM 33N and then transform the information to ETRF2000? Or the other way around?
    • Any other tips to help me out please??

Thank you SO much to any kind soul who have the patience to help a newbie

2

@IoshicDr1 I can only answer part of your query.

For me, the native GCP interface of WebODM never works. I always go to this interface it’s much nicer. Remember that you are only using that interface to create a file that you will then load with your images into whatever instance of WebODM you are using. Note also in that interface that you need to pick the EPSG code for the CRS that you used to collect your GCP points.

WebODM has a number of other idiosyncracies…but they are for another day.

4

Hi @IoshicDr1,

Welcome to our Community Forum.

It sounds like you’re going on the right path with using GCPs and your RS2, but a few important details about coordinate systems and transformations might be causing the issues you’re seeing in your orthophoto.

The main thing to check is the coordinate system your NTRIP provider uses, which is ETRF2000. This system is a bit different from WGS84, and that difference could be the reason why your GCPs aren’t aligning correctly with your imagery. Right now, you’re collecting GCPs in WGS84 UTM 33N, but your NTRIP provider is working with ETRF2000, and this mismatch could explain the stretched results you’re seeing.

To resolve this, you can try setting your RS2 to collect data in ETRF2000 instead of WGS84. If ETRF2000 isn’t an option in the Emlid app, you can still collect your GCPs in WGS84 and then transform them into ETRF2000 afterward. There are tools like QGIS or other online transformation tools that can handle this conversion.

It’s also important to make sure your exported .csv file is in the right format for the software you’re using, such as WebODM. Double-check that both the GCP data and the software are using the same reference system, like ETRF2000, to ensure everything lines up correctly.

Lastly, don’t forget to verify the height values you’re using, as they need to be consistent with the model your NTRIP service uses. If they rely on a geoid model, like Roma40, you’ll need to adjust the height differences between ellipsoidal heights (from WGS84) and the geoid heights your NTRIP provider uses.

2 Likes
5

thank you Mike, indeed that other interface is MUCH better… will test it… Thanks!

6

Hi Victor,
Thank you SO much for this… I’m not a professional surveyor , and don’t intend to be or pretend to be! I only use my UAV as a visual inspection drone and I’m just intrigued by the RTK mapping.

"…It sounds like you’re going on the right path with using GCPs and your RS2, but a few important details about coordinate systems and transformations might be causing the issues you’re seeing in your orthophoto.

The main thing to check is the coordinate system your NTRIP provider uses, which is ETRF2000. This system is a bit different from WGS84, and that difference could be the reason why your GCPs aren’t aligning correctly with your imagery. Right now, you’re collecting GCPs in WGS84 UTM 33N, but your NTRIP provider is working with ETRF2000, and this mismatch could explain the stretched results you’re seeing…"

QUESTION = OK, understood that WGS84 isn’t the one to do.
Now I downloaded the Italian Ministry Official guide and it says that the ETRF2000 (based on ETRS89) is mandatory since 2011. Good. In Italy it’s renamed as “RDN2008”.
However, in the document, it specifies these ones in this “system”.
a. - RDN 2008 - EPSG:6704 - geocentric - x,y,z CS axes
b. - RDN 2008 - EPSG:6705 - 3d geographic -
c. - RDN 2008 - EPSG 6706 - 2d geographic -
d. - RDN 2008 / UTM zone 32N (N-E) - EPSG:6707 - Projected - Nord,Est
e. - RDN 2008 / UTM zone 33N (N-E) - EPSG:6708 - Projected - Nord,Est
f. - RDN 2008 / UTM zone 34N (N-E) - EPSG:6709 - Projected - Nord,Est
g. - RDN 2008 / Italy Zone (N-E) - EPSG:6875 - Projected - Nord,Est
h. - RDN 2008 / Zone 12 (N-E) - EPSG:6876 - Projected - Nord,Est

Which one should I put in my GNSS Emlid Reach for a 2d Orthophoto work?
I’m surely at a loss here…

"…To resolve this, you can try setting your RS2 to collect data in ETRF2000 instead of WGS84. If ETRF2000 isn’t an option in the ReachView 3 app, you can still collect your GCPs in WGS84 and then transform them into ETRF2000 afterward. There are tools like QGIS or other online transformation tools that can handle this conversion…

It’s also important to make sure your exported .csv file is in the right format for the software you’re using, such as WebODM. Double-check that both the GCP data and the software are using the same reference system, like ETRF2000, to ensure everything lines up correctly…

This is great to know! In fact the software I used WebODM , I didnd’t specify the datum there…

"…Lastly, don’t forget to verify the height values you’re using, as they need to be consistent with the model your NTRIP service uses. If they rely on a geoid model, like Roma40, you’ll need to adjust the height differences between ellipsoidal heights (from WGS84) and the geoid heights your NTRIP provider uses…

Great to know this: so if I choose the datum ETRF2000 for horizontal, for vertical I should insert the same “vertical datum” that my NTRIP RTK Provider uses? (I believe it’s ROMA40 as far as I can understand)

THANK you so much for this… learning a lot as a complete newbie in this… it’s all a great experience and adventure for me

1 Like
7

You’re welcome! For your 2D orthophoto work, it’s best to use the RDN2008 system, based on ETRF2000, as per local standards (Italy). The right coordinate system will depend on your specific area. For example, if you’re in UTM zone 33N, you’ll want to select EPSG:6708 in your Reach receiver.

Based on your case of being located in Italy, here a the suitable options per regions:

  • RDN 2008 / UTM zone 32N (EPSG:6707): This is suitable if your work is in the zone that covers part of northern Italy.
  • RDN 2008 / UTM zone 33N (EPSG:6708): This is typically used for central Italy and is the most likely choice based on what you mentioned (since you’re already using UTM 33N).
  • RDN 2008 / UTM zone 34N (EPSG:6709): This covers more southern areas.

For 2D orthophoto work, you’ll most likely be working with RDN 2008 / UTM zone 33N (EPSG:6708), as it aligns with your earlier usage of UTM 33N.

Just make sure to align the coordinate system with what your NTRIP provider is using and double-check your settings in Emlid Flow to avoid discrepancies. If the vertical data (heights) is needed, use the geoid model your NTRIP provider uses (likely Roma40).

Great to know this: so if I choose the datum ETRF2000 for horizontal, for vertical I should insert the same “vertical datum” that my NTRIP RTK Provider uses? (I believe it’s ROMA40 as far as I can understand)

Yes, if you choose ETRF2000 as your horizontal datum, you should also match the vertical datum to what your NTRIP RTK provider uses, likely Roma40. This ensures that both your horizontal and vertical coordinates align correctly.

2 Likes
8

Victor, THANK you so much… this is gold!
of course, if you ever come to Rome, Italy, you have a huge free italian lunch already paid by me + a guided tour for free as well))) , don’t forget it (and I’m not joking)

Will catch up in case
I’ll try to have some fun in testing everything now -

1 Like
9

Very informal note… if you use a national correction service the EPSG (RDN/UTM) doesn’t matter.
The national service uses the RDN, and corrections are sent this way.
So… if you set your projection to UTM32-33-34 (Epsg 32632-32633-32634) the system expect correction in WGS84, but you give the correction in RDN2008. Result is: while conceptually wrong, and Never ever to be used the result is the same.

Suggestions:

  • Check the same point with the two EPSG (UTM and RDN)
  • Possibly use a certified reference IGM point. In Lombardia they are available for free. In other areas I don’t know…

Cheers,
Massimo

10

Hi Massimo,
thanks for the very informative tips - very informative and interesting to me indeed.
In Tuscany I found the certified IGM points online easily - need to search for Lazio ones

11

If you have a couple close to you they should be more than enough to validate the combination of correction service and projection.
Cheers,
Massimo

13

Thank you Massimo!