Aeropoints vs low-cost RTK

[rafael1147]

We now are considering purchase of the equipment to install GCPs. We are torn between buying Propeller Aeropoints or Reach RS receivers. I've read a lot about both systems and at the moment I see the only obvious advantage of Aeropoints - it is really easy to use. Overall Reach RS seems to be the better option as we can get 2 modules and set as many accurate gcps as we need for the fraction of price for 10 aeropoints.

Did anyone here have the experience of working with mentioned systems?

It would be great to hear comments on this.

thanks!

 

[bluelight support]

We now are considering purchase of the equipment to install GCPs. We are torn between buying Propeller Aeropoints or Reach RS receivers. I've read a lot about both systems and at the moment I see the only obvious advantage of Aeropoints - it is really easy to use. Overall Reach RS seems to be the better option as we can get 2 modules and set as many accurate gcps as we need for the fraction of price for 10 aeropoints.

Did anyone here have the experience of working with mentioned systems?

It would be great to hear comments on this.

thanks!

If you have access to CORS or RINEX data then you can post process down to 1cm+, and you only need one Reach RS, if you need access to realtime - you probably need 2 - I am going to get 1 to try - apparently they are excellent, however bear in mind which ever you go for you will need a good 10-60mins in each location to get a good carrier signal to process from. I also think I need to do a lot of work to get the Reach working with RTKLIB which is not user friendly, however is the most powerful option (and a lot cheaper)

 

[timmydjr]

If you have access to CORS or RINEX data then you can post process down to 1cm+, and you only need one Reach RS, if you need access to realtime - you probably need 2 - I am going to get 1 to try - apparently they are excellent, however bear in mind which ever you go for you will need a good 10-60mins in each location to get a good carrier signal to process from. I also think I need to do a lot of work to get the Reach working with RTKLIB which is not user friendly, however is the most powerful option (and a lot cheaper)

Hey Bluelight, I have been looking all over for info on setting up my P4P to use RTK/PPK. I am completely new to mapping and trying to jump on the fast track to learning all this stuff. I have ordered one Reach RS which will be here in a couple of days. I have read alot of post on the Emlid forum from a guy in Australia and he has a Reach mounted on his P4 Pro and is writing the RTK data to the exif files. BUT, he is not telling anyone how he is doing this. It is my understanding that using Reach RS and then a Reach as a rover on the P4 (or any bird) will pretty much do away with the need for manual GCPs. The million dollar question right now is how to match up the precise info with the Exif data from the Phantom 4. Do you have any detailed info on setting up a P4 like this? I sure hope that soon I will find some good detailed setup instructions from A-Z for people like me. Before I go and order the Reach module I am hoping I can find some good details on how to set this all up. I also have an I2 so I would ideally setup both of them up with precision accuracy. Take care!

 

[bluelight support]

My Reach RS is ordered and will be here next week - I've also looked at a couple of Chinese RK GNSS system as well and am getting one of those as they reckon with a static point they can get 3mm horizontal and 12mm vertical with post processing (with only 15 mins of static data) - so it looks like them for ground control points, and the Reach RS linked to the reach on the UAV to get relative accuracy on site.

Even on the expensive eBee RTK and other 20-30k systems I've seen, they are only getting 5-10cm accuracy (x,y) anyway for relative, and even my Inspire and Phantom where getting 20-50cm accuracy without trying. (z was over 1m however, so an issue for point cloud production and DEM). For me and I think a lot of professionals, currently surveys completed with a UAV and complimented by RTK or PPK ground control points and report sheets, is the way to go.

Of course, this all depends on your application and location. If I had a side which was too dangerous Im sure a ground survey less one with 50cm accuracy without human risk is much preferred.

I have read alot of post on the Emlid forum from a guy in Australia and he has a Reach mounted on his P4 Pro and is writing the RTK data to the exif files. BUT, he is not telling anyone how he is doing this.

People are cautious about giving info - ive been learning and researching for a couple of years and am now back at university doing a masters in this area to get the background knowledge you need - mapping is a deceiving sector - people think that buying a phantom and paying x per month to Pix4d or DroneDeploy makes them mapping experts - well let me tell you it doesn't. The only way to do mapping is to invest heavily in equipment, computers, software and training. When you paying more for a graphics card to process your data than you do on an Inspire 2 you know this is an extremely specialist area. I can tell you from experience, the drone is the cheapest part of the system and only a tool to get the data. Its what you do with the data to qualify it for us. Good luck, mapping is a fantastic industry - but do it properly.

 

[pandew]

In my experience the only benefit to using aeropoints would be on a new site where you just want quick georefencing and probably just a one-off visit- if you are doing repeat measurements/imagery of a site you are better off laying out some semi-permanent or permanent targets (I use a combination of painted targets and checkerboard vinyl floor tiles) and hitting these with RTK GPS.

I am on a 60 hectare construction site( freeway interchange) and I'm generally flying it every 2 weeks or so- certainly wouldn't want to be having to lay out aeropoints every time I fly... instead I have about 50 or 60 targets that cover the site. A lot of these are located on my existing survey control marks (so are coordinated/ levelled with a total station), others are just GPS'ed in.

I can cover this site with 3 flights at 70m AGL and have done extensive testing and checking of my results vs existing survey pickups. Have proven I can get comparable vertical accuracy to traditional GPS surveys- walking and driving pickups for earthworks volumes (in the order of +/-50mm or so). The big difference is the speed of UAV pickup vs traditional survey and also the amount of data and detail that you achieve (hundreds of millions of points vs maybe a few thousand with a normal end-of-month survey pickup as example).

I had the Aeropoint people out here demonstrating their workflow but I really couldn't see any use for this type of setup on this type of site- also the vertical accuracy that I got from their point cloud wasn't that great (about +/- 200-300mm). I think this was probably due to the limited number of GCPs in use (they had 5 aeropoints laid out) hence there were quite large distances between targets. I find that when this is the case the vertical accuracy near to GCPs is quite good but away from GCPs it can vary significantly (makes sense since the z values are just being determined by software interpolation). With the number of GCPs I am using the distance between each target is relatively less and I thus get pretty good consistency with Z values all over the site. The other issue I have is that I would rather process images on site (Agisoft) rather than uploading to cloud- would take me longer to upload the images to a cloud processing site than for me just to process images to point cloud/mesh and orthophoto on site! (and I have better control over GCP placement in images).

I have got a pair of Reach RS receivers here and have played around with them a little bit- need to do a bit more playing when I get a chance... I have no reason to believe they would give me any worse x,y,z accuracy than my other RTK receivers.

pete

 

[M3aerial]

Hi, I'll chime in with my experience with Propeller Aeropoints. I haven't used the Reach RS so I can't comment on it, however, we've completed many surveys with our Aeropoints and I am quite happy with them. They are easy to lay out, collect and upload data; they are simple and have no battery issues since they are solar powered. If you're operating within available RINEX areas then they're ideal. We are very well covered around us in Winnipeg, and Southern Manitoba, and the United States in general is VERY well covered with RINEX data. Laying out the points takes around 20 minutes, as does collecting. They must be on the ground for 45 minutes minimum, but usually it takes longer than that to fly the fields we fly (for agriculture), and mine sites. Overall we are very pleased with our investment. They paid themselves off very quickly.

 

[M3aerial]

In my experience the only benefit to using aeropoints would be on a new site where you just want quick georefencing and probably just a one-off visit- if you are doing repeat measurements/imagery of a site you are better off laying out some semi-permanent or permanent targets (I use a combination of painted targets and checkerboard vinyl floor tiles) and hitting these with RTK GPS.

I am on a 60 hectare construction site( freeway interchange) and I'm generally flying it every 2 weeks or so- certainly wouldn't want to be having to lay out aeropoints every time I fly... instead I have about 50 or 60 targets that cover the site. A lot of these are located on my existing survey control marks (so are coordinated/ levelled with a total station), others are just GPS'ed in.

I can cover this site with 3 flights at 70m AGL and have done extensive testing and checking of my results vs existing survey pickups. Have proven I can get comparable vertical accuracy to traditional GPS surveys- walking and driving pickups for earthworks volumes (in the order of +/-50mm or so). The big difference is the speed of UAV pickup vs traditional survey and also the amount of data and detail that you achieve (hundreds of millions of points vs maybe a few thousand with a normal end-of-month survey pickup as example).

I had the Aeropoint people out here demonstrating their workflow but I really couldn't see any use for this type of setup on this type of site- also the vertical accuracy that I got from their point cloud wasn't that great (about +/- 200-300mm). I think this was probably due to the limited number of GCPs in use (they had 5 aeropoints laid out) hence there were quite large distances between targets. I find that when this is the case the vertical accuracy near to GCPs is quite good but away from GCPs it can vary significantly (makes sense since the z values are just being determined by software interpolation). With the number of GCPs I am using the distance between each target is relatively less and I thus get pretty good consistency with Z values all over the site. The other issue I have is that I would rather process images on site (Agisoft) rather than uploading to cloud- would take me longer to upload the images to a cloud processing site than for me just to process images to point cloud/mesh and orthophoto on site! (and I have better control over GCP placement in images).

I have got a pair of Reach RS receivers here and have played around with them a little bit- need to do a bit more playing when I get a chance... I have no reason to believe they would give me any worse x,y,z accuracy than my other RTK receivers.

pete

Hey Pete, I wouldn’t say that Aeropoints can’t be used for the purpose you’ve identified. If you use a semi permanent GCP such as a board that is hammered down with a stake, and has an outline of where the aeropoint would go you’re good to go...

 

[bluelight support]

As an update - I have now complete gcp's with the Emlid Reach RS post processing with Effigis EZSurv and I managed 8mm horizontal accuracy with a static time of 10 mins per point. CORS stations were only 9 miles away so of course this can be worse depending on your data, and S3 sat triangulation files were also used . (OK I might be a bit biassed as we' re becoming an Effigis distributor for their software) - but of all the post processing software used, it was purely import one file and click one button, everything was downloaded and reported automatically.

 

[M3aerial]

As an update - I have now complete gcp's with the Emlid Reach RS post processing with Effigis EZSurv and I managed 8mm horizontal accuracy with a static time of 10 mins per point. CORS stations were only 9 miles away so of course this can be worse depending on your data, and S3 sat triangulation files were also used . (OK I might be a bit biassed as we' re becoming an Effigis distributor for their software) - but of all the post processing software used, it was purely import one file and click one button, everything was downloaded and reported automatically.

How large of an area were you covering?

 

[bluelight support]

How large of an area were you covering?

They were static point Rinex files used, so I was getting this accuracy for each ground control point.

It also depends on where your nearest base stations are as to what accuracy you could start with, but as a starting point I was very happy with the result.

 

[M3aerial]

Oh I see. Yeah the Aeropoints have super tight absolute accuracy on themselves as well. I think it’s 6 of one, half a dozen of the other.

 

[Dave Pitman]

I have read alot of post on the Emlid forum from a guy in Australia and he has a Reach mounted on his P4 Pro and is writing the RTK data to the exif files. BUT, he is not telling anyone how he is doing this.

Simon is selling his solution as a kit. I have just received mine but have not installed or given it a go yet. Reach on Phantom4

As an update - I have now complete gcp's with the Emlid Reach RS post processing with Effigis EZSurv

I am relatively new to rtk gnss but have been experimenting with 2 Reach modules which are exactly the same as the RS except without the enclosure, battery and LoRa radios. Perfect for PPK work though.

Reconciling the logs from the Reach has been fairly painless using the RTKLIB fork by Emlid. EZSurv looks really nice, but like much of software in this realm, is very expensive.

The Reach Module kit (a Reach board and Antenna) coupled with a power source and a target provides basically the same functionality as an Aeropoint at a cost of ~ $280 a piece + power pack and target (minimal $). In this role, they both rely on PPK with CORS. But the Reach modules can also be used stand-alone or in RTK mode which is not an option with the Aeropoints. The Aeropoint is an L1 L2 receiver so they are superior in that regard. But for PPK that's not as much of an issue.

For 10 Reach kits you would be into about $3000 compared to the $6000 for the Aeropoints. The Reach solution would not require any subscription to use them going forward. Not sure about the Aeropoints. In my area, a CORS network subscription is $1900 per year but the logs (for PPK) are free. YMMV

 

[Adam_Geomatics]

Hi All,

We have a new product for accurate aerial surveying without any ground targets. See https://www.veritypositioning.com/ for more details.

Thanks!

 

[vgo195]

As a surveyor I wouldn't rely on any RTK or PPK solution on the drone.
Check shots and also at least 1 GCP is needed for checking the position and accuracy of the model.
It is not a bad thing to have (rtk drone) but also geodetic receivers are needed for any precision job.
My 2 cents..

 

[Adam_Geomatics]

A good survey practice is to get couple of check points to see how accurate the model is, never trust on fully automated processing! This is clearly indicated in here Why Verity Positioning PPK - Geo Referencing System

 

[doktorinjh]

We have an eBee RTK and always use at least 5 GCPs and at least 10 check shots on a job. If you're going to be out in the field with an RTK unit and already getting check shots, then why not put down some GCPs or shoot some photo control? It's cheap insurance, you're already there, and it might save you in case you get a post-processing bust. This is one reason why the AeroPoints don't quite give us a lot of extra value; we already have the high-precision gear with us, so why do we need to add extra equipment? (I really like the APs and eventually want to get a set, but they're not feasible for our workflow.)

So, why do we have an RTK unit if we're running around and shooting GCPs and check shots? We like that we can use fewer GCPs and that they can have a less than ideal layout. For example, we survey transportation corridors and don't always have access to private land outside of that corridor. We can fly over it, but it might not be a good idea for us to walk through the private land and place aerial targets. With an RTK/PPK solution, we feel we can trust the data outside of the GCP controlled polygon more than without it. We've also only had the eBee RTK for one field season and we want to have a lot of repeatable data that shows the accuracies that we want before we start cutting out check shots and GCPs completely (if ever). There also may be projects where +- 1 foot may be acceptable and I think that we're comfortable with the RTK/PPK drone for that purpose.

With that said. I'm wary of hardware (or software) manufacturers that boast <0.2' vertical accuracies without ground control. And I'm even more wary of service providers that don't perform check shots to validate their surveys. At the end of the day, I want to be as accurate as I can, and giving up GCPs or check shots shouldn't be done in order to save a couple of hours of field time.

 

[Dave Pitman]

Hi All,

We have a new product for accurate aerial surveying without any ground targets. See Verity Positioning PPK - Geo-Referencing PPK System for more details.

Thanks!

Hi Adam,

your product looks nice and very similar in function and price to the Loki system. Just a comment though. The price you are asking is pretty high, IMHO. Maybe not when compared with what surveyor's have been willing to pay for gear. But with regard to uas systems. Perhaps your analysis is telling you that you will sell relatively few systems and that informs what the pricing needs to be. But consider that dji did not become the market leader with that model, and i'm not talking about the Mavic or Spark. The market is still open in this category and the first company that provides a quality system but is willing to "gamble" a bit on volume may just end up finding themselves with too much business.

If your target buyer is a surveyor that is moving into uas tools then maybe you have it right. If, however your target is uas service providers which at this point is a massive market by comparison, it may be worth some consideration.

Anyway, just my comments. Best of luck. It would be interesting to hear what others think?

 

[mdaemon]

I am quite sure, that real time RTK will not make GCPs useless. From the point of accuracy surveying, where you want to be absolutely sure, you will need them.
If sites are surveyed regularly, fixed GCPs will be the solution.
The amount in the photogrammetry is not that high if you have 10-15 GCPs (about 30 min).

regards
Martin

 

[Tiago Sá]

We now are considering purchase of the equipment to install GCPs. We are torn between buying Propeller Aeropoints or Reach RS receivers. I've read a lot about both systems and at the moment I see the only obvious advantage of Aeropoints - it is really easy to use. Overall Reach RS seems to be the better option as we can get 2 modules and set as many accurate gcps as we need for the fraction of price for 10 aeropoints.

Did anyone here have the experience of working with mentioned systems?

It would be great to hear comments on this.

thanks!

The main reason why I wouldn't recommend Aeropoints would be the price, but there's more reasons why you may consider buying an RTK system instead.

First, it might be a good idea to mark your GCPs with paint (or use features that are already in place) that way you won't have to recover the control points after the flight. In my case, sometimes I use rubber targets, but on two occasions they got stolen. It's sad, I know, but there's not much to do if there's no controlled access to the site.

Another important factor, is that sometimes you'll need to survey the same area multiple times, and it will be better to mark the GCPs permanently anyway. Finally, it's always useful to have some tool that allow you to bring your survey data back into the field.

I hope to be helpful.

Tiago

 

[Ataim]

I'm an RPLS in Texas, Oklahoma and Arkansas. We use Trimble R8, with network RTK. Much more expensive that either options listed, but we have the equipment for surveying first and drone second. This a typical control point that we will paint. And how they appear in our P4P from 150 feet. Easy to set and return to if necessary.