|
| Right now, most of the processing services do not give any feedback from a quality control point of view. Personally, I think that's immoral.
The most common inaccuracy you'll see is "striping", which is very obvious after you've seen the raw stitch, and usually comes from the lower end converted cameras. I suppose this is fine if you're aware of it and you're just using the map for a VISUAL tool, but it's dangerous to create management zones out of these outputs. I still have one of these outputs with striping up on our website and use it as our go-to example to start, because it's a good way to show the issues/concerns with remote sensing.
You can still use those cameras and services and get useful information out of the index maps, but I'd be very cautious about removing the human factor from the interpretation at this time. The automated corrections of inaccuracies is getting better week by week with just about every camera and processing service out there, but there are only a couple that I would trust right now.
If you want a "...see it for myself" scenario, try flying the same field at different times of the day or different weather conditions. Fly a field at noon when it's overcast, scattered clouds...fly it at 5pm when the shadows are longer. Now take those management zone shapefiles and compare. See what your application would look and how much those environmental conditions affect it, and if you got any kind of an indication from the service that the outputs should be used with caution.
What you're trying to do is where the money is with drones in ag (as opposed to just making pretty maps), but it's still a work in progress. Some companies are spending a lot of money on advertising, others on R&D. I'll toss my hat in with the R&D guys.
| |
|
PrecisionHawk/DJI Partnership Announcement
