RTKLIB with P4RTK

And yes, there is also the logistical part of it as doktorinjh pointed out. Where you can still be confident in the data based on fewer check shots and established PPK workflows instead of relying on poor GCP geometry due to access, etc.
 
The true advantage of PPK is the fact that the data will be more precision throughout the entire project. Processing with GCP will limit you to precise data only within the perimeter of the network of your control points. Data outside this area, and extreme high and low areas with no GCP's will be subject to error.

As a Land Surveyor, the PPK has proven to be a huge time saver. We always have check points from RTK ground locations. But traversing the perimeter takes time and often not possible.

I find it much better to use professional surveying software such as Topcon Magnet Office Tools to post-process the PPK data. When we fly the P4RTK we set up two Topcon Hiper VR GNSS receivers collecting static data at 5hz (0.2 seconds) epochs. I do use RTKLib software RTKconv to convert the 100_000x_PPKRAW.bin file from the P4RTK to a usable Rinex file format. Then post-process and add the GEOID with Magnet Office Tools. FYI the antenna height of the P4RTK is 192mm. I export Lat/Lon/OrthoHt data in CSV format (D.dddd, D.dddd, Meters).

I then use a software called GeoAdjust from WSI Software
This software will use the PPK data to create a corrected position of each image by loading CSV file, the MRK file, and the Images, no copy or pasting of data in Excel. The corrected CSV is then used in Pix4D or AgiSoft.
 
Last edited:
  • Like
Reactions: doktorinjh
The true advantage of PPK is the fact that the data will be more precision throughout the entire project. Processing with GCP will limit you to precise data only within the perimeter of the network of your control points. Data outside this area, and extreme high and low areas with no GCP's will be subject to error.

As a Land Surveyor, the PPK has proven to be a huge time saver. We always have check points from RTK ground locations. But traversing the perimeter takes time and often not possible.

I find it much better to use professional surveying software such as Topcon Magnet Office Tools to post-process the PPK data. When we fly the P4RTK we set up two Topcon Hiper VR GNSS receivers collecting static data at 5hz (0.2 seconds) epochs. I do use RTKLib software RTKconv to convert the 100_000x_PPKRAW.bin file from the P4RTK to a usable Rinex file format. Then post-process and add the GEOID with Magnet Office Tools. FYI the antenna height of the P4RTK is 192mm. I export Lat/Lon/OrthoHt data in CSV format (D.dddd, D.dddd, Meters).

I then use a software called GeoAdjust from WSI Software
This software will use the PPK data to create a corrected position of each image by loading CSV file, the MRK file, and the Images, no copy or pasting of data in Excel. The corrected CSV is then used in Pix4D or AgiSoft.

Thank you for sharing your workflow. Geoadjust seems great. I’ve been using Aerotas’ free excel file for processing the timestamp file with the corrected positions.

So you are using the PPKRAW file for post processing? How is that different from the obs file? I’m also not sure why you need the antenna height. Doesn’t Geoadjust do the lever arm correction to the camera?
 
Great discussion. I haven't quite decided to make the leap to RTK/PPK, but there are definitely pros and cons, and you guys bring up many good points to consider.

@amicron - You mentioned that you had been doing "several things wrong" in your processing workflow. Could you elaborate on which steps you changed?

@leegreen13 - I was about to hit the "BUY" button on the P4PRTK, but your workflow and multi-step conversions has frightened me. :eek: But really, thank you for explaining your process. I'm sure it gets easier after you've done it a time or two. What is the reason for setting up two receivers? Just redundancy in case one receiver fails?
 
DJI just released Terra software that appears to be their solution to PPK post-processing.

I use two Bases for a closed triangle to each drone point. This creates a mathematical closure and verification to ensure accuracy.

If you have a good NTRIP network and cell connection a the site, then RTK is seamless. No extra steps. The EXIF has the RTK corrected coordinates in it. Just go directly to Pix4D or AgiSoft. RTK is not as reliable as PPK. Also good to note, always collect static at your base, even when doing RTK. This way if RTK fails, you can still use PPK.
 
So you are using the PPKRAW file for post processing? How is that different from the obs file? I’m also not sure why you need the antenna height. Doesn’t Geoadjust do the lever arm correction to the camera?
For some reason the Rinex format of the obs file created by the P4RTK does not work with Magnet Tools (wrong format), also it does not contain the NAV file. RTKconv with create the proper RINEX3.02 format and NAV file.

Yes, GeoAdjust will calc the proper camera position after post-processing the PPK/Static data. However, the Post-Processing software computes a point-to-point position from the GPS antenna phase centers. Then using the antenna height we compute to the antenna reference point.
 
I still have many doubts about "everything".

Probably the only Phantom capable of offering a solid RTK / PPK solution is the expensive P4RTK. Maybe. In all the other models of Phantom 4, upgradeable with various kits that are on the net, there are too many errors and too many inaccuracies so that the final product is certainly correct.

At the moment, many multi-band GNSS solutions are coming out on the market at prices of a few hundred euros. I believe that this technology can be used to quickly create a network of ground points to which the postprocessing of photos can be linked.

Personally I remain convinced that having two independent sources, one of which is fixed to the ground, is still the best way to proceed.
 
However, the Post-Processing software computes a point-to-point position from the GPS antenna phase centers. Then using the antenna height we compute to the antenna reference point.

Antenna reference point? What is that? How does that relate to the camera/drone? Sorry, I just don’t understand.

DJI just released Terra software that appears to be their solution to PPK post-processing.

Where do you see that? I just looked through everything on their website and don’t see anything regarding post processing.
 
@amicron - You mentioned that you had been doing "several things wrong" in your processing workflow. Could you elaborate on which steps you changed?

1. I followed Aerotas instructions for CORS and did not select a CORS station close enough to work and then applied those instructions to processing my base.
2. Because of 1, I was not entering in the exact location of the base. I was “reading Rinex header.” I then used OPUS to get exact location.
3. I learned more about the options in RTKLIB. I downloaded the demo5 version from rtklibexplorer.com and read all about his explanation of the important settings. Very informative even though the majority was over my head. And the demo5 version comes with all his recommended settings. I also sent him my files and he confirmed I was getting good results.
4. Then, after I finally had RTKLIB giving me fixed results, I realized since I used OPUS to get the exact location of my base, I put in a 2m rod height out of habit and my elevations were way off. The location for post processing needs to be to phase center. I also downloaded the atx file for my base and use that but I can’t say for sure that makes a difference.

That’s pretty much it. Dumb mistakes or just not thinking.
 
  • Like
Reactions: Bigriver
I don’t think so. And Aeropoints is a waste of money IMO.

At the moment there is enormous technological progress in the field of GNSS reception. Today it is possible to buy a chip capable of managing the multiband (like the professional 30k ones) for a hundred dollars.

This technology can be used to measure the coordinates of a point on the ground in an immediate and extremely economical way. Post-processing technology has also become open source.

Phantom is a "closed system". Rather than trying to force existing products with external kits or buying a P4RTK (which has a thousand limitations due to proprietary firmware) I want to invest in "open systems" and on which I have full control.

With F9P chips you can build your network of intelligent GCPs for very little money ...
 
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
I’ve watched their videos and I’m not convinced they know what they are talking about. They are big believers in Aeropoints.
 
In my opinion the challenge is to find the way to measure GCPs quickly and economically.

In the near future, in Europe, "small drones" will be able to fly anywhere, even for professional purposes, without permits, licenses or courses. Drones weighed down by RTK systems, "strange antennas" or professional cameras will be a problem. No one (or almost) will use them anymore.

In the future I see small drones, with small cameras, and I see systems to make them work in combination with the ground control points in order to get the "survey grade".
 
Do you P4RTK users see any improvement in the quality of the point cloud and resulting orthomosaic over what is produced by just a P4P? Not from a georeferncing accuracy standpoint, but from an increase in quality (like less variations in point elevation differences, less noise, etc.). I understand that the cameras of the RTK model go through a more rigorous calibration process and wondered if it had any effect on the resulting model.
 
I am still wondering if the quality of the point cloud and resulting orthomosaic produced by the Phantom 4 RTK unit is any better than the Phantom 4 Pro. Has anyone compared the two?
 
So when DJI says that the camera on the RTK model is "better calibrated" (or something like that), is that sales hype or does it really produce better mapping?

As for point cloud accuracy, i suppose my question is, if you have a better geolocation of the drone when it captures the image, does that result in a higher vertical accuracy point cloud.? As in, do you have less variation in the elevations of adjacent points in the cloud? Do surfaces have less variation of elevation differences? I know this would be difficult to measure or quantify, but I am curious of the general opinion.
 
I find it very "suspected" that no one has compared the result obtained from a classic drone (with GCPs) with one obtained from an RTK drone (without GCPs). Maybe because the first solution continues to be better one ?

It would also be interesting to see how much to use an RTK drone with GCPs (spending a lot of money on the drone and a lot of time for the GCPs) improves accuracy compared to the "past".

Unfortunately on P4RTK we only have so much marketing and little real data...
 

Recent Posts

Members online

No members online now.

Forum statistics

Threads
143,066
Messages
1,467,352
Members
104,933
Latest member
mactechnic