Survey setup using DJI S900 w/ A2, Zenmuse Z15-A7 Gimbal, and Sony a6000 camera

[Peter Menet]

Hello Everyone!

I'm new very new to the GIS & drone world and I'm looking for some help/suggestions on coming up with an RTK solution for less than the $80k needed to buy a Sensefly eBee RTK. After much research, I'm planning to use the following setup:

• DJI S900 w/ A2 flight controller
• Zenmuse Z15-A7 Gimbal
• Sony a6000 Camera
• DJI Data Link (to use with DJI Ground station)
• DJI Lightbridge (2.4GHz for video downlink & transmitter amplifier)
• Futaba 14SG transmitter
• Switnav Piksi w/ 3DR 900MHz transmitter (for GPS datalog)

Although I much prefer the Ardupilot Mission Planner platform, I chose to go with the DJI A2 flight control because I wanted a stable, predictable, and well documented flight control system with built in loss-of-link and loss-of-motor failsafes. This is critcal because I'm working through getting the FAA's approval to operate commercially.

My plan is to use the DJI Ground Station's photogrammetry tool to create my flight grid over the area that needs to be surveyed and during post-flight processing, take the images from the a6000 and the GPS log from the Swiftnav Piksi and combine them to geo-reference the images. Essentially, I'm looking to use the piksi as a standalone (although possibly powered off the main hub) RTK GPS data logger. Pix4d has some instructions for how to do this using a Flytrex flight logger (https://support.pix4d.com/hc/en-us/articles/202560779) but the Flytrex does not support RTK.

My largest concern is accuracy. Since the datalog produced by the Piksi is done on a 10Hz interval I only get .csv information for every 1/10 secs. There is no direct interface between the camera and the Piksi so I can't be certain GPS data in the datalog is the exact time the shutter is triggered. If the drone is moving 10-15 meters per second having a discrepancy between the time the camera is triggered (time being established through a wifi connection and may only have 1Hz outputs) and the time in the Piksi datalog (established through GPS on a 10Hz frequency) could cause the geolocation to be off by significantly more than 3-5 cm that I would like.

I've heard of people using a Pixhawk connected directly to the Piksi and the camera to sync the times of the shutter and the GPS coordinate but I have no experience in that realm. My fall back plan is to simply use ground control points and don't bother with geo-referencing the photos at all but walking around laying GCPs is time consuming. (I'm beginning to understand why the eBee RTK costs so much...)

Does anyone have any suggestions? I know on the surface this should be a simple problem and I'm sure that my proposed solution will be much better than the 3-5 meters of accuracy using a standard GPS but like I said, I'm interested in the highest degree of accuracy possible.

Thank you in advance for your assistance. Feel free to contact me directly if you have any additional questions.

-Peter

http://menetaero.com

414-534-9980

[Peter Menet]

Also, I came across this little widget for the for the Pixhawk that will let the Pixhawk trigger the shutter. Since the Zenmuse is going to be facing only downward and the DJI Lightbridge will be plugged into the camera's mini-HDMI port, it leaves the multiport open to accommodate this device... Still have to figure out how to get a GPS shutter log...

[doug.a...@skydatapro.com]

Did you build this baby?

I use an s900 with the pixhawk and it works well.

[Anthony Kirk]

Let me enlighten you. We operate 20+ UAS. We started out with all the systems running the Dji A2 flight controller. Now we have one. All but one rig now rinning on the Pixhawk.

We have found the Dji A2 to be most un-reliable. Although the autopilot is open open source there is a LOT of very clever developers out there. One major advantage is, one a stable system is set up, if there is a problem the system WILL NOT ARM and fly off then fail crashing your rig.

We have had  many cached gigs running the A2 and never found a cause or solution. Dji are not interested in supporting there product. We also have 3x Dji Acc Waypoint systems costing £9,500 each in high end helicopters. The Acc Waypoint has many bugs and shortfalls but again Dji are not interested in supporting them. We replaced one with the Pixhawk and have had not issues at all with it.

Hope this helps.

regards

Anthony

[Brian Pfaehler]

S900 with pixhawk and piksi? I'm very interested in how you integrated the piksi and the camera...

--
You received this message because you are subscribed to the Google Groups "swiftnav-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swiftnav-discu...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Peter Menet]

Hey Brian, et. al.

So, we never got the Piksi to work. I had high hopes for it and had several people work on it but despite our best efforts, it turned out to be a dead end.

So, I’ve been working with a couple of different partners to come up with a purpose-built (vs. prosumer-adapted) GIS-specific platform. I can’t say too much about it but we’re hoping to have it hit the market in the next two months or so after some extensive field testing. What I can tell you is that it will be a Pixhawk driven system and will have a Sony a6000 payload with a “high accuracy GPS” hopefully capable of sub-10cm accuracy.

I’ll keep you all posted.

--

Peter A. Menet

http://menetaero.com

--
You received this message because you are subscribed to the Google Groups "swiftnav-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swiftnav-discu...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "swiftnav-discuss" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/swiftnav-discuss/IM9jYMAKZ6M/unsubscribe.
To unsubscribe from this group and all its topics, send an email to swiftnav-discu...@googlegroups.com.

[rai gohalwar]

Hi all,

There is a way to directly integrate the camera to the Piksi. By using the PC Sync Port (flash sync) on a camera, one can directly send a trigger command (level shift) to the piksi's debug port when the camera takes a picture. In Piksi that message is stored as an "External Event Trigger" and its TOW (GPS Time of Week) value is stored in that message. Using post-processing tools we are able to accurately interpolate the position at that TOW value using the neighboring position messages. This allows for the user to get accurate position tags of the images takes from the flight. The best part about this setup is that if the Piksi to Pixhawk Integration Guide is followed ; all the data is logged right on the Pixhawk. 

I hope this information helps. 

Rai 

To unsubscribe from this group and stop receiving emails from it, send an email to swiftnav-discuss+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Steve Zeets]

Hi Rai,

Do you have more information on how to wire the PC Sync Cable to the Piksi?  I also read your White Paper and couldn't find the information.  Also, I couldn't find the python scripts to post-process the data and create a file for Pix4D.
I have a PixHawk, Piksi, Hot Shoe, and everything else needed.

Thanks!
Steve

[dzo...@swift-nav.com]

Steve,

You should be able to wire the PC cable directly to the Debug Header on Piksi. The pin labeled "Debug0" is set-up as the event input pin.  You will also need to adjust the piksi firmware settings under the "external events" grouping.  We have a wiki page the describes how to use the external events here: http://docs.swiftnav.com/wiki/Timestamping_External_Events  .  I apologize that it is a little incomplete but it should get you started!

Best,

Dennis