Here where I work we don't take the drone into too much consideration when it comes the quality of the products that we can provide, granted our sUAS is very high end. We have only been using a fixed wing aircraft made by Trimble that is specifically designed for surveying. We know that the quality of our products depends on how it is processed and gathered. We use survey grade GNSS receivers or conventional total stations to set out GCPS and to observe the more critical areas of a survey when it comes to elevation, especially when it needs to match some existing grade somewhere. The camera onboard is also very high quality, not just a "UHD Camera".
We use the Trimble UX5 HP which logs a static GNSS vector as it flies and must be run with a local static base station logging it's own Data. Trimble claims that their processing Software (TBC) paired with this particular drone eliminates the need for GCPs through the use of Post Processed Kinematic methods that they derived.
Trimble is well known to the surveying field to be high end surveying equipment and is probably blindly trusted when they make these statements. We, however have found through our own experiments that it is simply not true. I have come across horizontal differences of more than a foot and vertically even more when comparing a processed flight with survey grade GCPs to one without them.
This, a survey grade company (Trimble), still doesn't yield the accuracy needed to rely on for any type of design without the use of survey grade GNSS receivers and the knowledge to understand the shortcomings of the photogrammetry performed. This, in my mind does not bode well for new services such as PX4d to be reliable for survey grade data yielded from a point cloud they generate. I agree that they can take a point cloud and make it useful but to rely on it for photogrammetry scares me. We cautiously use TBC and understand that in order for a survey to receive our stamp that it must be controlled well.
We not only use our system for Topographic surveys but for Alta surveys as well as supplemental data. Deriving the feature line work and most of the surface (if requested) is done along with our survey technicians.
It is clear to me that the data that can be gathered by sUAS can be very accurate if processed correctly, but that it does not eliminate the need for ground work, nor the need for a licensed surveyor to produce and examine the data.
I agree with R Martin above that surveying is still a very profitable profession to enter, having just entered it myself as a 24 year old LSIT. The need for surveyors is growing very quickly as well in New Mexico.
Someone said above that it would be foolish not to get into the newest technology available as a surveyor and I couldn't agree more. Surveyors for the most part are not young people and do not like to have to learn to work with new technology. That is where we young surveyors can help the previous generation while they help us to understand how to determine a boundary, and how to use all of the equipment they have been accustomed to. The Idea of GNSS was rejected fairly widely and thought of only as a fad, those that believed this were missing out on increasing profitability in a lot of projects. Don't let this technology pass you up as a fad. It is a great tool that we use frequently and it will only continue to grow and improve.
Sorry if I seemed to jump around.