Mavic Air 2 compatability

What is the compatability of the Mavic Air 2 with open drone map? Is it capable of doing stockpile estimations by itself, or is some sort of correction data necessary?

1 Like


OpenDroneMap isn’t a flight planner (yet?), so on light of that, I’m going to be answering a slightly different question than what you asked.

If you have a flight plan with sufficient sidelap and overlap, OpenDroneMap will absolutely stitch your collected pictures into usable data.

As for stockpile analysis, having GCPs will help a ton with Z accuracy, which is critical for volumetric analysis.

I personally don’t do analysis within WebODM (I use QGIS out of habit/convenience for me), so I can’t speak to the volumetric analysis workflow within WebODM except that it does exist.

If you can, show us what you are working on!

Yeah, that is what I meant. I am somewhat new to this whole world, and so would the gcps also need gps data associated with them? And how would a GCP on the ground help with z accuracy? would you have a set height of the gps module?

New is good because I think this all is a load of fun, but there is a not insignificant learning curve ahead, haha. Buckle up!

Yes, GCPs are both in your imagery (ideally) as well as surveyed point data. If you’re marking them, go big. Less headaches later. If you can’t mark, make sure you pick easily distinguishable features from the sky.

Yes, most times you’ll need to compensate for the height of the GPS module. Many have an adjustment in the menus somewhere for exactly this, oftentimes tied to a particular part-number/model of pole/mount. If not, you can usually set an offset. Failing that, your waist is usually about 3ft off the ground or so and you can manually offset the values later (or you can measure boot sole to belt-buckle and get your true offset, and just hold the GPS module exactly at your belt for each point collection).

GCPs help because GPS units in general are frankly quite awful at Z accuracy. Having GCPs taken from known elevations (if you’re lucky enough to have survey monumentation on-site) or with RTK/PPK or other survey-grade tools will reduce your Z error from well in excess of 10m (typically) to fractions of a centimeter for RTK/PPK.

If you don’t have access to those tools, you can use GPS Averaging with a really long baseline (collect time) to approach meter-grade accuracy with most phones. GPS Point is my favorite Android App for this. Even entry-level Garmin eTrex units can do Averaging, so most dedicated hardware units can as well.

1 Like

okay, cool, that makes sense. When it comes to PPK, how is the data that the base station is getting being used to correct the drone data without being actively connected to the drone? It doesn’t seem very intuitive, although maybe it’s more complicated than can be explained here. The way I believe it works is the base station and the drone are collecting the information about the satelites position and such, but the base station is much more accurate, and so the data that the base station collects replaces the data that the drone collects during post processing. is this correct?

1 Like

PPK is great because you don’t need to worry about in-situ internet access to get corrections in real time like RTK. Just need a GPS that writes out the data properly.

The correction stations are all over the US and operating constantly. Your job is to pick the closest one to get the best corrections when you PPK your data.

Your GPS data (provided they are writing out RAW RINEX sentences) are all that are needed to do PPK since you’re comparing and adjusting the recorded data against the stations reference/corrections. You have the freedom to PPK (or not) your data pretty much whenever you feel like it.

It isn’t replacing your measurements (how could it? It isn’t where you are), it is being used to augment your measurements to reduce systematic errors.

I like to think of it as 90% of the benefits of RTK with about 10% of the fuss/expense.

Though to be fair, rtklib is not the most friendly software suite ever :rofl: