rtk / dgps - Has anyone pushed towards decimetre accuracy

For everything that doesn't fit into the other (sub-) forums

rtk / dgps - Has anyone pushed towards decimetre accuracy

Postby Mattpgreen » Thu Dec 04, 2014 10:50 pm

Hi all
I've looked through the forums but not found any discussion about trying for best possible location data.
Has anyone looked at using rtklib in any way and/or setting up a second gps at the GCS for dgps?

I'm ramping up to get an M4 to FC a sub 100g photosurvey platform and I'm aspiring to best possible camera location data.
Mattpgreen
 
Posts: 48
Joined: Tue Oct 07, 2014 11:06 am
Location: Melbourne, Australia

Re: rtk / dgps - Has anyone pushed towards decimetre accurac

Postby LPR » Fri Dec 05, 2014 2:18 am

Matt

Welcome.
I can't help but wonder what you mean by sub 100g photosurvey platform? With 100g a multicopter with not be able to fly very long or high. The camera will need to be a couple of grams. I would think you might be able to photosurvey a football field on a calm day but not much more.

There has been some discussion about RTKLIB and Swifts Navigation's Pixsi RTKGPS system.

http://swiftnav.com/index.html

DIY Drone has some people attempting to get the Pixsi working with the pixhawk autopilot but does not seem to be helping their accuracy so far.

Larry
LPR
 
Posts: 1323
Joined: Tue Jun 26, 2012 1:29 pm
Location: MN, USA

Re: rtk / dgps - Has anyone pushed towards decimetre accurac

Postby Mattpgreen » Fri Dec 05, 2014 3:13 am

Hi Larry
yes to the survey only being a football field (or a few) - per battery on a calm day and yes to the camera only being a few grams ... and this can still satisfy many potential job specs.
I stalked the swiftnav and although it is clearly a way forward, I'm trying to find a way of getting close'ish to their output with lower cost components.
Can you please guide me to the RTKLIB discussions that you refer to?
Thanks, Matt

ps i see that ublox are working towards comms (lte/2G/3G/WiFi/BT) GNSS and "Inertials" in a single chip - that's interesting.
Mattpgreen
 
Posts: 48
Joined: Tue Oct 07, 2014 11:06 am
Location: Melbourne, Australia

Re: rtk / dgps - Has anyone pushed towards decimetre accurac

Postby Kisssys » Fri Dec 05, 2014 3:18 pm

Matt

The code can log the necessary RTK data for post processing. I have done it on RC racing aircraft to overcome the g limit that the normal gps receiver imposes. You would need a second gps receiver and live in an area that can receive the corrected data to do what you want and it's possible.

Let's say you were getting position data that you can calibrate to 50 cm accuracy, I would want to know if I can deal with just a few things that come to mind.

How accurate would the gimbal have to be to point the camera with that accuracy?

If you didn't stop their do you think the optics on the small camera's would fulfill those requirement, The ones I have light enough wouldn't do it?

If you get gps updates 15 times a second how fast can the copter be going and how much shutter delay error can you tolerate to stay within 50 cm.

Cheers
Steve
Steve
Kisssys
Kisssys
 
Posts: 1340
Joined: Sat Jun 23, 2012 9:23 pm

Re: rtk / dgps - Has anyone pushed towards decimetre accurac

Postby Mattpgreen » Mon Dec 08, 2014 9:41 pm

Hi Steve
I'm using Agisoft PhotoScan for the survey output - the more accurate the camera position the better but in fact no camera position data is workable as well ... just more processing time.
A several MP sensor is enough to build the models....of course I'm working towards a 2015 plan to incorporate the emerging >12MP smart phone sensors.
In addition, I'm looking to integrate visual odometry such as - https://www.youtube.com/watch?v=2YnIMfw6bJY - this has very significant navigation and survey output processing time reduction prospects.
Mattpgreen
 
Posts: 48
Joined: Tue Oct 07, 2014 11:06 am
Location: Melbourne, Australia


Return to General Discussion

Who is online

Users browsing this forum: No registered users and 20 guests