-

[veDtdvCGL$Db.hxd]

-

 

[flightmike]

Possible loss of some gps. Watch your transmitter bars on the controller and when they start to drop that's far enough.
If you put anything between the controller and the MP you will lose reception.

 

[Mako79]

Yes. you will lose sats. Keep them above 8 counts (though its GPS+Glonass, you need min of 7 GPS or Min of 7 of Glonass and NOT combined).
Also watch out for high wind drafts.

 

[geigy]

Good point. Position errors near a cliff could cause a crash as the Mavic decides it is suddenly in the wrong place. I would have collision avoidance on, and face toward the cliff, so that if a position error occurs that would cause the Mavic to "compensate" and ram into the cliff (and climbers!) you will have an extra safety cushion.

Also, killjoy drone police would suggest that you only go close (< 150 ft) to climbers that you know personally, but of course I'm not that type of person.

 

[geigy]

Sounds like an interesting project. One thing to look at, if you haven't already, is the various mapping softwares out there that can compile a 3D model of a structure. These are used to quantitatively measure progress on construction projects and similar applications.

Here's one article that compares a couple of these packages. Usually you can do a few projects for free.
DroneDeploy VERSUS Pix4D | Heliguy

 

[geigy]

Wow - now there is an application!

Re parallax, yes, that's how these 3D reconstruction programs work. Another buzzword is auto-correlation..

These algorithms have been around for decades and are fairly accurate. I imagine they could be very valuable for your application. There are glitches. For example, repeated patterns in the cliff texture, and protruding branches and roots might fool the software. Also, these are generally designed for horizontal mapping, not vertical, though some of these programs are used to reconstruct radio towers and the like.