dronexsurvey Posted at 8-8 09:25
Would be good to set up base over known point for long time ; but how long the Base station can run is limited by WB37 battery life - only lasts about 45 minutes in DJI Base station in our experience so can't log or calibrate longer than that, might not be long enough for some applications.

The D-RTK 2 can be powered by an external battery or AC power adapter. You're not limited to a single WB37.

That being said, we still generally site in our base location with another RTK GPS with which we can log at least 15 minutes of data, spit out a RINEX file, and send to OPUS for rapid-static or static processing. We use the coordinates from the OPUS report to enter into the D-RTK 2's setup for when we fly.
Haven't been able to figure out how to do that with the D-RTK 2 only.

patiam Posted at 2-14 10:05
Anyone else using the P4RTK want to chime in?

(bump)

Hi There,
The interface is ordinary for this issue. I do a lot of work in remote jungle areas and I setup the base  let it settle down and find itself-monitor the altitude readout untill it is stable- then take a pic of lat long altitude and manually put it into DJI.
Ordinary at best. I check co-ords using another GPS and take an averaged reading over 15 minutes (minimum 12 sats in view) as a check that I'm not metres out.
This is not geodetic  precision and I dont pretend it to be.
My results are acceptable for the work I'm doing around the 150mm+/_ .
Another thing I found useful is extending base station another 1.5 metres by adding a sleeve for the dji tube. This helps in scrubby areas for sat view and also Comms.-see pic below.
DJI please add an averaging function for the rtk base and then an option to put in antenna height and then accept that as the position.cheers Myke

patiam Posted at 4-15 11:19
Just thought I'd update this thread since it has been some time since I reported on our continuing effort to develop a workflow and test the accuracy of the P4P RTK. We've got enough flights in to get a better feeling for how the system performs.

Unfortunately, while we have continued to enjoy great (1-3 cm) XY precision and accuracy, that is not at all the case for Z.  As @ewischropp reported, within-flight vertical precision is quite good (that is, for a given flight elevations are internally consistent). But between flight precision and accuracy are unacceptable. Doesn't matter whether we use NTRIP or the D-RTK 2, we're seeing up to 0.8m DSM elevation departures between consecutive identical flights, and up to 1.3m differences between the DSM and known surveyed Z at our checkpoints.

@patiam,

Thanks for sharing all of your results. I just purchased a Phantom RTK. I probably would not have had I read this comment beforehand.  I use drones to calculate stockpile volume. Z is the most critical component when you are running a volume against a known base. .8m inaccuracy across 10-15 acres leads to unacceptably larger errors.  I  can't think of a reason why you would see different results from the same flight plan flown at different times.

We might be able to eliminate a few of the variables and help isolate the problem with a PPK solution. Have you tried PPK on this data set?  I am far from an expert on PPK but could process the data for you if you are unfamiliar.  I would love to help solve the mystery. Until this problem is identified and resolved, I see no place in my drone inventory for an RTK Phantom.  I

@Skyveyor112,

Thanks for the comments. Please see the Accuracy and Precision of P4P RTK thread for more in-depth discussion and analysis of the limitations and solutions for this platform.
The issue seems to be with optics, not positioning. Many have found improved results by tweaking the FL of the camera and or doing a camera calibration. The system is capable of producing quite accurate results, just not at all altitudes right out of the box. The other issues with the kit are that DJI touted the D-RTK 2 as having other operation modes and that still has yet to really materialize almost a year after its release, as well as software design flaws that are incompatible with survey workflows (geofencing, requiring sign-in, poorly implemented failsafes, poor UI)

I've not used PPK processing but again the most serious issue I've seen is with photogrammetric processing of the imagery, which seems to be addressable in post. YMMV, but don't despair that you made a poor choice in buying a P4RTK.

patiam Posted at 9-9 08:34
@Skyveyor112,

Thanks for the comments. Please see the Accuracy and Precision of P4P RTK thread for more in-depth discussion and analysis of the limitations and solutions for this platform.

@patiam,

Thanks for the steer. Read with great interest and a bit of hope.

patiam Posted at 9-9 08:34
@Skyveyor112,

Thanks for the comments. Please see the Accuracy and Precision of P4P RTK thread for more in-depth discussion and analysis of the limitations and solutions for this platform.

First off, I would like to say that I have found this and numerous other threads regarding the workflow for the P4 RTK and processing (especially in Pix4D) extremely helpful and informative, so thanks very much for all of this information.

I have a question for you patiam (and anyone else that might have some wisdom to share of course).

This sort of goes right back to the beginning of this discussion about how to setup the D-RTK 2 base station. My thought is, rather than adjusting the height of the D-RTK 2 in the field to account for the antenna height (1.802m seems to be the consensus), is there any detriment in doing this in post processing?

For example, in Pix4D (and I'm sure in other software packages also) there is the ability to adjust the coordinate system of the images being imported (just like adjusting the output coordinate/height datums). Could you not just add the antenna height here, like this:

I'm very new to this, but to me this would be easier than having to input the base station height everytime you setup.

dodger_rog Posted at 9-10 21:23
First off, I would like to say that I have found this and numerous other threads regarding the workflow for the P4 RTK and processing (especially in Pix4D) extremely helpful and informative, so thanks very much for all of this information.

I have a question for you patiam (and anyone else that might have some wisdom to share of course).

dodger_rog,

I don't have factual or empirical data to support my position but think what you are suggesting would work. I think this is the same as adding 1,802m to every photo, which is what adding 1.802 to the base elevation does.

That aside, I try to give the system the best information I can from the outset and would input the correction as I enter lat, long and elevation for the base position. Should only take a few seconds to add 1.802 to known base elevation.

I'd do some testing to compare the results.

Good luck and let us know what you discover.

Skyveyor112 Posted at 9-11 04:40
dodger_rog,

I don't have factual or empirical data to support my position but think what you are suggesting would work. I think this is the same as adding 1,802m to every photo, which is what adding 1.802 to the base elevation does.

Another thing to consider is that someone else might be processing the imagery, so you may want to have all the corrected geolocation data written to each image as opposed to doing in post. We often get contracted to be data collectors, though we certainly love to fully process our data all the way through if possible. Just a thought.

Skyveyor112 Posted at 9-11 04:40
dodger_rog,

I don't have factual or empirical data to support my position but think what you are suggesting would work. I think this is the same as adding 1,802m to every photo, which is what adding 1.802 to the base elevation does.

Thanks Skyveyor112,

Will be doing a large survey later this week, will let you know how it goes....

EcoSolutionUAV Posted at 9-11 06:37
Another thing to consider is that someone else might be processing the imagery, so you may want to have all the corrected geolocation data written to each image as opposed to doing in post. We often get contracted to be data collectors, though we certainly love to fully process our data all the way through if possible. Just a thought.

That's true, although in our case I'm the one going to be doing the processing, at the moment that won't affect us. However it's good to remember for in the future.

In theory you should be able to enter the antenna offset at any time in the workflow. We've steered away from doing it later  as we were attempting to nail down the systematic accuracy and precision issues and wanted to account for every offset possible during acquisition.

EcoSolution's point about processing documentation and metadata is a good one.

patiam Posted at 7-19 08:47
You enter the known coordinates of the base station Under RTK Settings>Advanced Settings>GNSS Coordinate Input. As you noted, you must include the antenna phase center height above the point on the gound in the altitude (Z) you enter.

Hi Patiam,

I am a little confuse when you enter the coordinates of he base station to the GNSS Coordinate Input. When you enter the altitude coordinate you have reduce the antenna height? or what you have to do.

@djiuser_vncBzCtirKOB-

I haven't seen the new firmware so somebody who has can correct me if this has changed- but yes, when you enter the altitude of your D- RTK 2, DJI suggests that you include the height of the antenna phase center above  of the benchmark over which you are set up (so you enter the benchmark altitude + the antenna height). Many more established software applications for setting up RTK base stations include a separate field in which you can enter (and store) the antenna height of your tripod so that you can simply enter the Z of the benchmark without having to do the math yourself for every new setup. Not so with GS-RTK, at least as originally released.

The fact is that the antenna height offset can be introduced at any point, but as has been discussed here and elsewhere, there are few good reasons NOT to account for it during acquisition and plenty of reasons why it IS  a good idea to do so.

Hello guys,

Just read all 3 pages, very good thread here. I'm a surveyor here in Philippines and we recently bought this P4RTK and was wondering if the OLD method of using multiple GCP's would still be doable with good results using P4D?

There is an upcoming project and we have not tested out this P4RTK tbh so i dont have the luxury of time to redo the survey if ever i don't use GCP and proceed with the survey and get unacceptable discrepancies.

I'm curious if I use the DRTK Base station and still use multiple GCP's and primarily just process the whole dataset using the old method of using multiple GCP's.

With the old processing method considered is it okay to leave the DRTK base altogether and just use the P4RTK just like P4 and P4Pro and use multiple GCP's? Just due to time constraint of this work.

Safe flight everyone

@Zyrandiel-

Yes of course you can still use GCPs as you would with a non-RTK P4. In fact it is recommended.

@Patiam

Thanks for the quick response mate =D

I tried a short test flight yesterday and found that you can DISABLE the RTK in the settings.

After disabling RTK the display in the GS RTK app now showed GNSS and i was able to fly the drone.

Is it safe to assume that surveying in this setup would just be like using a standard P4/P4Pro with help of multiple GCP's?

Safe flight everyone

@Zyrandiel-
You're vey welcome!

Yes, flying w/o RTK the P4PRTK should perform as well as a "regular" P4P (perhaps better, due to the improved time stamping). But unless you are just trying to simplify things or have less kit to lug around, I think you'll have improved results using the D-RTK 2 as well and operating in RTK mode, even if your base station location isn't known to a high degree of accuracy. If it were me, and I had no means of getting an exact benchmark location (and no NTRIP data source were available), I'd set the D-RTK 2 up at my site well in advance (like 15 minutes or more) and let it run until the position stabilized before doing my flight(s) with RTK enabled. You might not enjoy the full potential accuracy that RTK can bring but it will almost certainly be better than autonomous GPS.

Tell me, how are you getting your GCP positions? If you have a way to get high accuracy positions for your GCPs you should be able to do the same for the location where you set up the D-RTK 2...

But the answer to your original question remains yes, without RTK the P4P RTK should be as good as a regular P4P.

@patiam

Thanks for your time and response. I truly appreciate it.

Yes i will utilize the D-RTK base station once I have more time. Work sched is kinda tight this month.  I want to have a full operation test using D-RTK before I use it for a project.


For your other question:

I did a 3-hour static observation using TopCon GNSS receivers. I established 2 control stations simultaneously using this method on site.

I uploaded the RINEX data to PPP Canada GNSS Precise Post Processing webpage after a day to get a RAPID solution output.

I can use the established 2 control stations to do an  RTK localization calibration and take observations over multiple GCP's on site.



I have a different question which involves Pix4D.

In our work, we mainly focus on bathymetry and other oceanographic stuff.
The only time we deal with land survey is if we have a client that wants us to survey the land portion which is adjacent to the bathymetry area.

We mainly use a Tide Gauge unit installed on site to record tide levels during the whole survey project.

We tie up Tide Gauge level to our project control stations on land usually in Mean Lower Low Water Datum (MLLW).

Now our project control stations will have values in MLLW.

If I setup my D-RTK base station over one of my control stations on site how would I setup my vertical datum when i process in Pix4D?

I've read you guys here use Geoid models and ellipsoidal values. I'm curious how would i relate my whole survey to the MLLW datum that is in my control station.


Safe flight everyone

Zyrandiel Posted at 10-22 04:54
@patiam

Thanks for your time and response. I truly appreciate it.

@Zyrandiel-
Thanks for describing your GCP methodology- that'll work

Funny that you should bring up the topic of bathymetric mapping- before getting all droney I spent >20 years doing hydrographic mapping using MBES, LiDAR and other tools. Not sure if you ever heard of the California State Waters Mapping Program, but that was us.

We got away from tidal datums long ago, and it was one of the best things we ever did. Using PPK to produce a tightly-coupled trajectory from our POS-M/V data yields a vertical accuracy of a couple cm, so we just map to the ellipsoid and then get to NAVD88 using a geoid model (for example GEOID 12B). In the old days, even when we had in situ tide gauges we would have to play with tide modeling over large areas and there were always nagging artifacts. Even worse with predicted tides. Besides that, MLLW is a moving target over time and space, so it makes more sense to tie everything to a more "reliable" vertical datum, especially when doing bathy/topo and in light of sea level rise.

But back to your question, if you need a way to get your topo stuff into MLLW you can always use something like VDATUM to go from MLLW to NAVD88 or vice versa. Unfortunately Pix4D only allows a single adjustment factor per project to be applied to your ellipsoid-based altitudes in your images, so if yoour area is large and the offset varies too much you may have to brak up the project, or just keep everything in HAE and apply the variable adjustment to the output products in GIS or elsewhere.

Cheers!

Hi,

I recently started using the RTK Base Station and came across the same issue regarding the antenna height. I would be interested if there is any update. Has the value of +1.802 m proven to be correct?

After 20 or 30 projects, right now my process is very simple... i let the drtk-2 to use the coordinates it gets, i dont care about it's height. In fact im using another tripod (a real one) with a tribrach and an adaptation. Then i use at least 1 GCP (could be near de base station), but usually i use 2 or 3 (or more depending of size)... and that's all. This procedure give me the best results in precision and time consuming.

In my first projects i measured de antenna height, get its coordinate with an RTK GPS, and fix the DRTK-2 position with lat / long / height .... waste of time.... even if you fix the base station with 1mm precision the results of an orthophoto (tested in PIX4D and Agisoft Metashape) just SUCK

Sorry to hear you aren't seeing good results when entering accurate coordinates on the D-RTK 2... Have you tried the camera model Focal Length tweaking discussed in this thread?

https://forum.dji.com/forum.php? ... 711&pid=1908175

Hi Guys

Sorry i havnt gotten back here sooner.

Have any of you good folks out here used Mavic 2 Pro for Drone Mapping?

P4RTK is out of stock here in Philippines and we need at least 2 Drones next month thats why i am wondering f Mavic 2 Pro is any good for Drone Mapping

Any info would be greatly appreciated.

Fly safe! =)

Zyrandiel Posted at 10-20 21:13
@Patiam

Thanks for the quick response mate =D

Zyrandiel,

Sorry I didn't see your post and reply sooner.

Be careful flying the PRTK in GNSS mode. Some of my testing suggests the PRTK doesn't correct for non-standard altimeter settings (29.92 in the states).  If the altimeter is anything other than 29.92, the drone will not fly at the altitude you programmed. If the altimeter is high, it will fly 10ft higher for each 0.01 change in the altimeter setting. We have seen the drone fly several hundred ft above planned altitude on high barometric pressure days. We haven't flown on any low pressure days, but I expect the same behavior in the opposite direction. The implications of which are two-fold; 1) safety of flight. You may run into terrain, trees or other obstacles. 2) You will have gaps in imagery coverage because the flight was planned at a higher altitude.

I posted my results in the terrain following thread.
https://forum.dji.com/forum.php? ... d=192850#pid2000105

Good luck,

Dave

patiam Posted at 2019-7-26 11:02
Glad to help. Let us know how it goes.

Your avatar- Clemson?

Sorry - I haven't been on the forum in awhile, and I don't get notices that someone has replied!  Yep, Clemson Tigers fan.  Tough game Monday night, but at least it was to a bunch of good guys with a phenomenal quarterback and a coach with a personality that rivals Dabo's!

patiam Posted at 2019-7-19 13:42
Sure! I'm familiar with Aeropoints but have never used them... Do they give you the Z value for the ground underneath their center, or the top surface of the unit (I'm guessing the latter, given that you're supposed to use them as GCPs)? If its for the upper surface you'll want to account for it (subtract the thickness of the unit) or have the unit in place and put the pointy end of the D-RTK 2 pole on top of it...

Patiam,
Re your question about exactly where the Z value for AeroPoints is calculated, you're right - it's the top surface.  So to get the ground height at that point, they recommend you subtract their thickness of 32mm.  Thanks for the reminder of that.

Skyveyor112 Posted at 2019-12-29 05:53
Zyrandiel,

Sorry I didn't see your post and reply sooner.

Hey buddy,

Thanks for this information.

Appreciate it.

Regards,
Ryan

Finally, some breathing space

https://ww.dronedj.com/2020/01/08/dji-phantom-4-pro-v2-0-is-finally-back/#

Hi guys, I bought a Phantom 4 RTK DRTK-2. I read all pages about all issues with the base station. Thanks a lot for all informations!!
Just to be sure that I understand all well (my english is pretty bad, I'm from France).
If I use a dongle and a NTRIP in my base station, with a height of 1,802m. In mode 3, I can take the right position of my GCP. Let say 5 targets. I place one more for set my position for the base station.
After that, I place my tripod and my base station in the middle of the last one. I go in mode 1, and in RTK setting, I enter manually the position of this target, XYZ. Is it work? Should I calculate the Z value plus the 1,802m?One more thing, Propeller ask me what is the datum vertical geodesic from the base station. Is it height ellipsoid or geoid model?
Thanks for your help

Yes to everything you said, although I have not gotten Mode 3 to work (haven't tried a lot).
Regarding your last question, you said Propellor, are you using AeroPoints?

Your P4RTK wants the DRTK-2 elevation to be in HAE, as all your bird positions will be in HAE. As for your GCPs, depending on your workflow and processing software, you can potentially have them in any coordinate system/xyz datum you wish, as long as your software supports that.

patiam Posted at 2-3 10:13
Yes to everything you said, although I have not gotten Mode 3 to work (haven't tried a lot).
Regarding your last question, you said Propellor, are you using AeroPoints?

Hi Patiam, no, I'm not using aeropoint because I use the DRTK 2 instead of it. You right, I can choose system datum with Propeller. That's why I asked it because I didn't know which system datum DRTK2 use.
Thanks a lot for your help

Hello Everybody,
Thank you for sharing your experience with Phantom RTK. It is really helpful.
I just started to used and did test flight. Didn't get good result yet with Dronedeploy.

My Z error is about 40cm. I cant figure out why.
When I rocess it in PIX4d I've got better result.

Any help appreciated.

Sur_veyor Posted at 3-10 08:52
Hello Everybody,
Thank you for sharing your experience with Phantom RTK. It is really helpful.
I just started to used and did test flight. Didn't get good result yet with Dronedeploy.

What altitude did you fly? A brief description of your processing workflow would be helpful as well.

Your Pix4D results  are in the expected range (<= 5 cm Z RMS), especially if flying > 40m AGL and not tweaking Focal Length of camera model.

If DD is giving strongly different results, then there is either something you're doing differently there or the software is not quite right...

patiam Posted at 3-10 09:12
What altitude did you fly? A brief description of your processing workflow would be helpful as well.

Your Pix4D results  are in the expected range ( 40m AGL and not tweaking Focal Length of camera model.

My workflow :
-50m AGL
-Setup GCP's using GPS Topcon with accuracy 2cm . CS is NAD 83 3TM  M114- local.
-Re-calculate D-RTK base XYZ to WGS84. I keep my ortho elevation because there is no option in DD choose ellipsoid. Anyway from DD tech support  I've been told to do that.
-add 1.802m rod height to D-RTK base elevation
-setup D-RTK base on CP
-drone flight with fix position
-upload images to DD and tag checkpoints(No 3d GCP have been used for processing)

PIX4d :
the same images, no ellipsoid have been used.
tag checkpoints(No 3d GCP have been used for processing)

Sur_veyor Posted at 3-10 09:44
My workflow :
-50m AGL
-Setup GCP's using GPS Topcon with accuracy 2cm . CS is NAD 83 3TM  M114- local.

PS I have only evaluation version of PIX4D and all project we process in DD

That difference you got (40cm) is very close to a constant im applying to the base height.
After 20 or 30 project i got an average of 48cm.

My workflow:
-70m AGL
-Setup GCP using GPS Spectra SP60 with accuracy 2cm in WGS84
-In one of these GCP is mounted the D-RTK base station, measured its height
-Setup DRTK with known coordinates. Lat & Long goes directly from GPS.
-With the height i got from gps i do this: let's suppose i got 91.50m ground elevation where is mounted and 2.01m base height then: 91.50m + 2.01 - 0.48m = 93.03m (and that's the height i enter in the controler)
-drone flight with fix position
-upload images to Agisoft Metashape (no GCP are used for processing)

guillermomeiner Posted at 3-10 10:12
That difference you got (40cm) is very close to a constant im applying to the base height.
After 20 or 30 project i got an average of 48cm.

Thank you for sharing that info. It is really weird. I need to figure out that constant offset theoretically :-) I am waiting what Dronedeploy tech support will say. It is so annoying.

Hello
Glad to chime in after having read so many times these treads.
I use Agisoft and just had a pretty good result using P4RTK + DRTK2 on a known point.
My system is RGR92 UTM 40S on Reunion island. We have a local geoid that i am unable to use except on a local software for a single point.
Quick process:
- get the calibration of the P4RTK with an oblique flight and plenty of gcps.
- put the drtk2 on a known point. Had to convert the xy to wgs84 dd but i kept my geoid altitude just adding the 1.802m of the pole.
- Nadir flight at 70m
- In agisoft i declared flying at 70m and put the confidence of pictures of xy at 3cm and z at 1m (which is the average difference with the ellipsoid where i flew)
_Re use the calibration parameters as a known value
- I ended up with accuracy below 5 cm without gcps and below 3 cm using them in the reconstruction process.

guillermomeiner Posted at 3-10 10:12
That difference you got (40cm) is very close to a constant im applying to the base height.
After 20 or 30 project i got an average of 48cm.

What I figured out after some research:


In my case Separation between Ellipsoid WGS84(DD) and GRS80(NAD 83) is -0.431m in that point.
This is close to your offset.

guillermomeiner Posted at 3-10 10:12
That difference you got (40cm) is very close to a constant im applying to the base height.
After 20 or 30 project i got an average of 48cm.

Hi guillermomeiner,
Why is your base hight 2.01m? Are you not using standart rod(included with D-RTK)?