Thread Tools
Sep 07, 2016, 01:57 PM
Registered User

Wow!


Looks like a great idea! I've been using a couple of reach units for post processing, but nowhere near a turnkey solution. I'm definitely in for a set! We all need more accuracy....less pretty maps : )
Sign up now
to remove ads between posts
Sep 07, 2016, 10:29 PM
Registered User
Looks awesome Steve. Looking forward to hearing more about this.

Al
Sep 08, 2016, 09:10 AM
Registered User
Squawk's Avatar
I would like to know if it's still worth buy a +5K D-RTK System for our DJI Matrice 600..

Especially sine it only seems to be 5 Hz and we need to fly relatively fast to survey large areas.
Sep 08, 2016, 10:59 AM
Registered User
Thread OP
Hi Squawk, I want to be careful how I respond to this because it is not my intent to speak badly of others products.

However I am all for factual discussions!!!

So here it goes. From what I see M600 is going to get you good in flight positional holding, repeat-ability for flight paths, and no magnetic interference using DRTK. For some that is useful but for survey it is more important to locate the camera position IN FLIGHT to as accurate as you can. Think of it like having a GCP at every camera position in the sky.

It does not look to me like they are headed for a direct geo-reference solution at this time. Yes it can map but you will still need ground control. 5hz is only one of the issues they have to solve.

You can only fool your clients for so long with reports they don't understand and maps that look nice. Companies are starting to PENG stamp and sign these reports and they need proper data to back those deliverables up.

The report submitted in the m600 thread shows camera positioning errors of 8 feet horizontal and 80 feet vertical. The ground control corrected this but the point is not to use ground control and why are the camera positions out so far? I can speculate quite a few reasons.

Our clients almost always do their own check shots and compare them to our data without us knowing. Now that is a check shot



Anyway here is the kicker for our clients when we start selling Stryder and that is CUSTOMER SUPPORT.

Our direct geo system will be SEPARATE from the flight controller meaning if there is an issue with the data we can fix it. I could not imagine asking DJI tech support to correct an issue I am having with my survey. That would be hilarious.

Next there is currently no way to set up over a control point with the DJI system. That means no repeatable data. Yes you can enter in co-ordinates for your DJI base but how is that thing going onto a tribrach to center over a point? It's not.

So far from what I see DRTK is not a direct geo solution and it may not ever be their intent. Surveying and supporting surveying is an entirely different beast when done properly.

We want to educate and help people create accurate data and grow with us.

There is a reason nobody buys chinesse survey equipment. The world rely's on Topcon, Trimble, Novatel etc because it's reliable, your data is worth money when it's accurate.

Lets add to that you are stuck with DJI camera's which are decent but if you want to fly a 24 or 42mp full frame sony....

We are processing some recent data from Stryder that we are very excited about We won't divulge until we have done full testing and re-testing though.

BTW Stryder uses the A3 system and E1200 motors (Stryder was done lone before the wind1 came out) for flight so it's still super easy and you can use the same drone to do inspections with all of DJI's cameras.

Don't get me started on carrying around 18 batteries and all the chargers for three flights, and the argument about taking them on a plane is insanity. They are not letting you on a plane with 6 of those batteries let alone a case of 18 no matter what the sticker says for watt hours.

We and others like us usually ship the batteries via ground two days before. Easy.


The end goal is that clients or service providers are not flying around the country because they will either have their own drone locally or will hire a local for the work.

That is my opinion on the state of things right now.

Steve
Last edited by saabguyspg; Sep 08, 2016 at 12:39 PM.
Sep 08, 2016, 11:03 AM
Registered User
Thread OP
All that being said we are not ready to start selling Stryder or OBS yet. We are gathering feedback from people via email, phone and on here.

I imagine we will be on the market at the same time as DRTK.

Yes I know you can buy DRTK right now but currently you don't know what you are getting so I don't consider that a released product.

Already we are adjusting and changing our plan to better suit suggestions we have heard.

Thank you to those who have emailed and called.

Steve
Sep 08, 2016, 12:03 PM
Registered User
Squawk's Avatar
Thank you for your extensive explanation!

I will read over it a few times to make sure I understand it and when not, to know what question I should ask.

For now though, I have 1 question, is Stryder a complete drone?

That will not be for us, as we've gone the M600 route and are going through the (expensive) process of getting it certified for commercial use.
Batteries are a thing, but I do like the integration and cleanness.
It is to be hoped though, that DJI will come out with their Multi-charger soon..

Anyway, can we utilize the M600 with a Stryder system?
Or would OBS be our only option when using that platform?

Thanks for your time, really appreciate it!

Koen
Sep 08, 2016, 12:14 PM
Registered User
Thread OP
Hi Koen, if M600 and DRTK don't provide a geo referenced solution locating camera positions to <5cm accuracy in the air then you will need ground control of some sort (OBS would work or any survey grade equipment to locate your GCP)

Stryder is a complete system yes. Drone, case batteries etc.

Keep us in mind for your next system and if I can help with your M600 data I will try to .

Steve
Sep 08, 2016, 12:23 PM
Registered User
Thread OP
OBS

So I have received some emails and phone calls regarding OBS.

It sounds like a lot of people would perfer to pay more for OBS if they could reduce the wait time at each GCP down to 5 minutes or less. With single frequency you have to wait 15 minutes to be safe at each GCP. This is fine for a lot of people but we are seeing there may be two markets.

The issue here is increase in cost because we would need to use dual frequency receivers and antenna's.

We are sourcing some just released dual frequency hardware and will do a cost estimate.

I think we may run two Kickstarter campaigns, one for single frequency and one for Dual.

As a high level estimate I think the dual frequency system will be <5,000 USD but you will be able to pick up points 3X faster so you can use it for large sites.

Any thoughts guys?

Steve
Last edited by saabguyspg; Sep 08, 2016 at 12:36 PM.
Sep 08, 2016, 01:31 PM
Registered User
Thread OP
Here is a great write up by drone deploy explaining the need for gcp or direct geo

http://support.dronedeploy.com/docs/accuracy

Steve
Sep 08, 2016, 03:50 PM
Registered User
Squawk's Avatar
Thanks again for your reply Steve.

Yes, will keep your Strider project on my watch list. But first we really have to earn some investment back before spending for a new drone with certification etc.

Just as a 'in between solution'.. Could we not just use the M600 with RTK system and place that in each corner of the area, plus the centre, and leave it there for a few seconds (5Hz RTK GPS), record the positions and use that for GCP's?
Sep 08, 2016, 04:00 PM
Registered User
Thread OP
Theoretically yes you could do this however what you want is to locate your GCP as accuratly as possible.

Your GCP should be <5cm accurate, we try and get mm accuracy on ours... usually 1-2cm is average

In order to use your drone as a static GPS observation device you will need to know where the phase center of the antennae it is currently using (there are two on m600) exactly in comparison to the center of the GCP target.

As you can now imagine setting the drone over a point on the ground and measuring within a few mm so that you have no error added to the GPS error is impossible even if you knew what antennae is being used at the time.

Now if you sat over your target on the ground and took a photo assuming they have the lever arm correction in their software (they will never tell you) then you could take the position of that photograph XYZ from the EXIF and subtract the distance from the camera to the ground.

But again.. how do you get the camera perfetly over the target? Maybe by lining it up in the live HD view... but this is a very inaccurate way to do this... but fun to talk about.

We were going to release a Jig with Stryder to allow you to do this very thing... however it's better to use the right tool for the job so we invented OBS

So the short answer is that I feel this to be too inaccurate of a solution.

Does that help?

Steve

Quote:
Originally Posted by Squawk
Thanks again for your reply Steve.

Yes, will keep your Strider project on my watch list. But first we really have to earn some investment back before spending for a new drone with certification etc.

Just as a 'in between solution'.. Could we not just use the M600 with RTK system and place that in each corner of the area, plus the centre, and leave it there for a few seconds (5Hz RTK GPS), record the positions and use that for GCP's?
Sep 08, 2016, 05:34 PM
Registered User
Thread OP
oh and you need to place GCP based on required end accuracy.

As a rule of thumb you are good with 150-200m max between GCP.

One GCP should be added at the highest point and one at the lowest point on your site.

Then you place your check shots BETWEEN your GCP. As you move further away from your GCP the data will degrade.

Data is always good near a GCP and the GCP location.

Hence why checking next to a GCP is pointless... so to speak

Steve

Quote:
Originally Posted by Squawk
Thanks again for your reply Steve.

Yes, will keep your Strider project on my watch list. But first we really have to earn some investment back before spending for a new drone with certification etc.

Just as a 'in between solution'.. Could we not just use the M600 with RTK system and place that in each corner of the area, plus the centre, and leave it there for a few seconds (5Hz RTK GPS), record the positions and use that for GCP's?
Sep 08, 2016, 05:53 PM
Registered User
Squawk's Avatar
Thanks Steve!

The second post I understand.

But the first post..
The antenna used will have a different LED color than the backup antenna on the M600.
So one could position the 'master' GPS antenna over a CGP (let's say a yellow-black half-circle) that you would leave there when flying, right?

Then of course, as your second post explained, you'd have to do another set of 4 check GPC points between the GCP's that you don't use in the mapping software, but later to check for errors as far away from the GCP's as possible.

Hope I get this right?

Of course this would be easier with 5 sets of your cubes, but they won't be available until next year.

Btw, would 6 cube sets not be better to do the highest and lowest GCP as well, since they will hardly ever be the same as the 4 corner points or the centre point of the area..
Sep 08, 2016, 06:26 PM
Registered User
Thread OP
Hi, even though you will know what antennae is used you will have to put the camera over the target. I would assume DJI has input a lever arm correction between the camera and the GPS. This correction you won't be able to modify.

When you take a photo I am assuming the GPS position is of the camera not the GPS itself.

But if you place the camera over the target then you have to line it up with the target, measure down from the lens and then change this as an offset in your GPS position.

All of this is fun and it will likely help your model but honestly I would not use this as a method if you had to provide accurate deliverables.

Sorry our OBS won't be ready for you

With reagards to the 5 or 6 GCP's I think with our new pricing and hardware sourcing we will just make two products.

1) Single frequency OBS that you have to wait 15 minutes at each GCP location. This will be 950 USD

2) A dual frequency OBS unit that you will wait max 5 minutes and possibly down to about 3 minutes. That will be about the same price as a 6 piece kit. And I am loosely estimating about $4,000 USD for the kit? that is pretty high level estimate there but it has to be under 5K IMHO

Oh btw 5 pieces would work because the base is a target edit. sorry I understand what you mean now. You would fly the survey then grab one OBS and go do your check points after because they still would work as individual rovers.

Hope this helps.


Quote:
Originally Posted by Squawk
Thanks Steve!

The second post I understand.

But the first post..
The antenna used will have a different LED color than the backup antenna on the M600.
So one could position the 'master' GPS antenna over a CGP (let's say a yellow-black half-circle) that you would leave there when flying, right?

Then of course, as your second post explained, you'd have to do another set of 4 check GPC points between the GCP's that you don't use in the mapping software, but later to check for errors as far away from the GCP's as possible.

Hope I get this right?

Of course this would be easier with 5 sets of your cubes, but they won't be available until next year.

Btw, would 6 cube sets not be better to do the highest and lowest GCP as well, since they will hardly ever be the same as the 4 corner points or the centre point of the area..
Sep 08, 2016, 07:43 PM
BirdsEyeView Aero
ErichF's Avatar
This looks like a similar concept to Propeller Aero's AeroPoints system.


Quick Reply
Message:

Thread Tools