AeroPoints: new smart ground control points for easy survey grade drone data

Hi all,

At PropellerAero we think it is too hard and expensive for commercial drone operators to capture survey grade data - we needed a solution which was separate from drone or camera, and didn't require expensive GNSS rovers out on site.

Today we are launching our new AeroPoints - easy to identify ground control targets with built-in PPK to capture very accurate (~2cm in X,Y, 3cm in Z) positions to georeference your aerial surveys.

AeroPoint ground control point by Propeller

The units are fully solar powered, water-resistant and last weeks on the inbuilt LiFePO4 Battery.

We have been working on the concept for the last 18 months - and the accuracy we are getting is pretty exciting:

The AeroPoints correct off local RTK basestations, and we currently have coverage across AU, USA and the UK, with more countries coming online shortly. In areas outside RTK correction zones, the AeroPoints use a complex mesh algorithm (like PPP) to correct the position data, giving <1cm internal accuracies and decimeter-grade global accuracy after a few hours of recording.

This mesh algorithm gives our AeroPoints much greater usable baseline lengths than traditional, instantaneous RTK as well - as we measure the same point from more distant base stations the difference from the Topcon control remains tolerable:

Post survey, the AeroPoints look for a wifi hotspot (SSID: "propeller", password "propeller") and will connect automatically to the Propeller servers and upload all their information, where we then tie against local CORS data and generate highly accurate points.

For those who already use the Propeller platform for processing and visualisation, the AeroPoints will automatically tie down your data - and for those who use desktop processing (pix4d/agisoft), we output a PDF report of qualities and a CSV block ready to go straight into your workflow.

A set of 12 runs for $6000 US, good for 150Ha (370acre).

Really looking forward to improving data accuracy across the industry: coarsely georeferenced information being used for technical applications is particularly frustrating for us at Propeller!

Cheers,

Rory

More information at https://www.propelleraero.com/aeropoints

Hi All,

I have +20 years of photogrammetry and drone mapping experience and respectfully I don't believe that this technology would help in cost and time saving neither in achieving reliable results, due to the following reasons:

Targets have to be physically placed on the ground before flight and collected again after each flight (as opposed to traditional targets which are permanently placed or marked on the ground).

Site access restrictions, difficult and dangerous terrains makes this solution to fail at first instance similar to traditional ground targets. (For these targets you must visit the site twice to collect them back)

Assuming, the above mentioned issues are not a problem for some users, dependent of the area size number of ground targets (and therefore price) has to be increased to warranty the reliable mapping accuracy. This is a known argument in photogrammetry that using GCPs for non-metric cameras only provides good results at proximity of GCPs and in between GCPs accuracy still is determined by the image coordinates (from low accuracy GPS). You may have noticed this in some of your projects that even getting tens of GCPs doesn't provide satisfactory results and specially the height accuracy could be off by meters for homogenise textured grounds (water, wheat, crops, trees, ...)

For post processing and extracting target coordinates, raw data has to be processed against a known CORS network data which normally are available after 24 hours, so for even a small job one has to wait 24 hours to get the coordinates.

You need to be a member of diydrones to add comments!

Join diydrones

Comments

Aami October 4, 2016 at 3:02pm
Hi All,

I have +20 years of photogrammetry and drone mapping experience and respectfully I don't believe that this technology would help in cost and time saving neither in achieving reliable results, due to the following reasons:
- Targets have to be physically placed on the ground before flight and collected again after each flight (as opposed to traditional targets which are permanently placed or marked on the ground).
- Site access restrictions, difficult and dangerous terrains makes this solution to fail at first instance similar to traditional ground targets. (For these targets you must visit the site twice to collect them back)
- Assuming, the above mentioned issues are not a problem for some users, dependent of the area size number of ground targets (and therefore price) has to be increased to warranty the reliable mapping accuracy. This is a known argument in photogrammetry that using GCPs for non-metric cameras only provides good results at proximity of GCPs and in between GCPs accuracy still is determined by the image coordinates (from low accuracy GPS). You may have noticed this in some of your projects that even getting tens of GCPs doesn't provide satisfactory results and specially the height accuracy could be off by meters for homogenise textured grounds (water, wheat, crops, trees, ...)
- For post processing and extracting target coordinates, raw data has to be processed against a known CORS network data which normally are available after 24 hours, so for even a small job one has to wait 24 hours to get the coordinates.
Sobido September 1, 2016 at 3:37am

Hi Rory

Good idea to build a simple CGP system. However I have the same feeling than Tom .

I guess that you build a raw datalogger and then YOU post-treat the signal with available Ntrip element (Why you mention certain country only ? ) around the survey, via a software like RTKLIB or somewhere like it. Not really new in fact.

The business model is not necessary in the RAW signal catching process but in the following post treatment and the total price / CGP could be then comparable with a "simple" pro RTK survey .

I do also agree with you that a onboard RTK solution is not yet perfect ( I tested the ublox C94_M8P for example but with some disappointment ) and we do need still some CGP . However, because of the progress on the following onboard GPS generation ( M6<M8N<M8P float) , we can slightly minimize their number ( clearly better "Track" response).
Rory San Miguel August 30, 2016 at 12:40am

Hi John, actually prices are in USD. We always recommend 8 AeroPoints now for surveys. 5 is bare minimum on a small site but with 8 you really start getting cm grade even between points - really exciting for us.

We really like the Emlid Reach - fantastic team and well built product. In practice, it is quite hard to make this work in reality: even the best implementations give us 20cm accurate data. In saying that, definitely better than coarse geotags so by all means!

Cheers,

Rory
John Campen August 27, 2016 at 9:34pm

I assume your prices are in AUD on your web site and not USD? Might be wise to make that clear if they are?
I am also wondering that even for small areas you recommend 8 units at least. Which is $4000 right? In contrast I am using Emild Reach that cost around $600, and getting up to 1cm accuracy covering around 9km2 (900ha) What are the pros and cons with AeroPoints against Emlid Reach? One con I can see is all the walking around to set-out the GCP units?
Moderator

Gary Mortimer August 26, 2016 at 3:06am
Tom August 25, 2016 at 11:26pm

Are those the popular m8n+data logger+PPP, so i imagine we upload the "collected" data from these pods and they do the post processing for us with their base station access in those countries? 500$ each is a reasonable price IMHO for the pod assembly & hardware but as noted not sure if renting a rtk system is cheaper and simpler compared to this workflow.
Moderator

Gary Mortimer August 25, 2016 at 1:40pm

I'm just about to talk to Rory on air about these boards http://youtu.be/AtmdxNZgW-Y I will ask him Errol
Errol Burrow August 25, 2016 at 1:29pm

This is great but I have a couple questions:

1. what is meant by the pricing model exactly.. "A of 12 runs for $6000 US, good for 150Ha (370acre)." has confusing wording, is this saying that the user is charged $6000.00 to process up to 370 acres? That sounds a little pricey to me.

2. what local RTK basestations is the system compatible with?
Moderator

MarioSpeedwagon August 25, 2016 at 9:46am

Love it. This is one of those ideas where I always think "It makes too much sense for someone to not already be working on it".
Pascal P. August 25, 2016 at 1:01am

Lol ! Agree with Luke. You are lucky here if your disposable target is still on your pictures during the survey. But good idea for closed areas like mines or private properties.

of 2

This reply was deleted.