My thoughts for the weekend,
I have recently attended the TSPS convention and Tech EXPO. Trimble seemed to have a crowd around their new UAV platform that claims to be able to gather DEM of 5 cm accuracy. WOW. The FAA is expressing great concern about safety Liability and privacy concerning UAV's. As a Private Pilot myself, I would hate to run into a UAV that was not registered or had a not filed flight plan. I know that our field crews have forgotten things way too many times when heading out to the field, such as datacollectors, files, hammers, rodmen, etc. I would not trust them to file a flight plan just to get a quick job done. I think that a certification program needs to be looked at to operate UAV's. On the other hand I would love to get my hands on one!
I got to thinking about ways to take photos from the ground level while the camera is attached to the range pole and getting a RTK shot on the PNEZ of the camera position. Yesterday I was thinking about it while surveying at a construction site.
How can I produce a DEM from the combination of a 20mp point and shoot Nikon mounted on my RTK rover setup?
I am always trying to figure out how to do things in the DIY mode. I took a total of 4 shots with the camera at about 20 different points around a pile of dirt. I spent the evening and most of this morning trying out different software that could be downloaded for free. I got some results from VisualSFM that look promising but I am still working on the process and steps to make anything look like a usable point cloud yet. PLUS, this is a huge learning curve!
I am wondering if there are any others who have wondered about getting Volumes from ground based photogrammetry methods. I would think if such a method could be devised, then it might be a good check and confirmation of the accuracies of volume measurement from RTK or/and Total station or/and scanning methods.
There seems to be a way "in Theory" to import raster images and scale to known points and certain angles and draw the breaklines. Autodesk Infrastructure Modeler seems to be the best option but there is not very many tutorials that I can follow to achieve the goal.
Any Thoughts From the forum?
I think incorporating something like the DJI phantom would work very in the situation you described. Take a bunch of photos at ground level, and then add in a few from above. They make an attachment that allows the users to tilt and pan the camera, and see what the camera is seeing. Might need an hexa or octo-copter to run that, though. The payload of a DJI phantom isn't great. I would like to see the FAA open up airspace to professionals with UAVs, if we keep them below 200' and a certain distance (1 mile?) from airports...
> I think incorporating something like the DJI phantom would work very in the situation you described. Take a bunch of photos at ground level, and then add in a few from above. They make an attachment that allows the users to tilt and pan the camera, and see what the camera is seeing. Might need an hexa or octo-copter to run that, though. The payload of a DJI phantom isn't great. I would like to see the FAA open up airspace to professionals with UAVs, if we keep them below 200' and a certain distance (1 mile?) from airports...
I agree that there should be limits set for the different elevations/certificates. Most UAVs for Mine DEM would be under 500 feet if guessing. I would be concerned with the preprogrammed flights of a UAV that has the ability to hit 13,000 ft asl being used by a nonprofessional.
What I am looking for is a total ground based approach that could be used for cross checking volume surfaces using photos from a "point and Shoot" camera. I would not want to send a field crew out with a $ 1500 camera setup. The new "point and shoots" are claiming almost no distortion in the lens these days.
Could a 25' topo rod be incorporated to have the camera at a higher angle? The canno sx260 has the ability to be set to shoot continuously and is under the $300 limit.
The nikon that I used was a $90 Nikon s3500 with a 64Gb card. Total was about $150 bucks. I bent a piece of all thread and screwed the camera to it and mounted it using plumming pipe straps, $15 bucks at home depot and I have enough to make three more. the camera lens was about 1' below the base of the GPS receiver. So I have the camera position coordinate accurate to about 0.2 . Will the camera position matter in the software? or does it simply calc the position of the camera by the matches in pixels?
Ranger,
Don't get me wrong, I am all for UAV's in the hands of professionals. I am just looking for the ground approach.
Thanks for your reply,
Wes
I would recommend the www.diydrones.com web site to pick-up some basic information.
The AMA writes insurance for rc modelers under 400 feet. Above that, the FAA takes over.
There is a bunch of bad information that the surveying magazines have put out there about
the $10K FAA fine. The average surveyor can run a rc model plane for under 400 feet and
get his work accomplished.
But the bottom line is they can't do it legally at this point. They may get away with it, but the law is fairly clear isn't it? You can't use a UAV commercially, yet. Despite what Trimble wants you to think.
I think it's odd that people will follow the FCC narrow-banding requirements so closely but ignore the FAA rules because this is a new cool business opportunity. Trimble is especially guilty of this, will no longer service my 5603 robot with old-style radio and site the FCC laws, but advertise the heck out of an UAV that would be illegal to use.
Trimble told us that at this point only government agencies are able to operate UAVs in the US and each flight requires a separate authorization. I haven't heard them tell anyone to operate outside FAA rules.
And yet three.rivers says you can do it just fine, Trimble is causing that confusion. Trimble may not be telling you to break the law, but they're marketing the heck out of it and making the tools. They are at very least encouraging the behavior in hopes that once legal they will be first in line.
I wonder if you gave them a credit card number if they would tell you no..?
RTK, Volumes, Ground based PICS. Thoughts?
mmmmmm.....
Lets say I have no ability in the world of RC devices.
Lets say I don't want to spend a year learning to fly RC's to the point where I dare velcro a pricey camera to it and then fly it in a work space where my actions will be critiqued by my clients.
Let's say I want complete ground based system to incorporate RTK, Pictures taken from the ground, and Volume calcs for under $500.
Let's say I want a cheap check on my volume check and maybe have a tool where I could cheaply derive volumes with greater accuracy.
just saying....
RTK, Volumes, Ground based PICS. Thoughts?
Oblique photography is harder to control, but if you placed targets in the images and surveyed them in.. at the top of pile, bottom, sides, etc. I think it's doable. You'd have to have good contrast and sharp images, that's harder to do than you'd expect. You'll want a wide angle and fast lens for good coverage and proper exposure. I think you'll find that repeatable results are difficult.. you'll find weird bubbles and anomalies in the top of the pile (where the camera can't fully see) those results will throw your quantities way off.
This is why an aerial-based is best. I've done some testing with this, but I don't do any volumes work and never seen the high-quality results you're envisioning.
RTK, Volumes, Ground based PICS. Thoughts?
I don't know what it costs but try the Stock Pile APP for the IPhone.
RTK, Volumes, Ground based PICS. Thoughts?
I do not know the specific language in their private contracts with the FAA.
BP Amoco, Royal Dutch Shell, and Conoco Philips all have licenses.
Some of the applications include marine life monitoring for barge and pipeline
routing, ice thickness measurement and surveillance for bringing oil from Russia
to small Alaska towns and boroughs, and existing pipeline surveillance.
Supposedly the University of Alaska is hiring surveyors for location of various
types of refuges.
I can tell you with absolute certainty that Trimble will NOT sell the UAVs in the US to anyone who does not have the proper permissions to own and operate it. If you look at their web site, they don't even list a US distribution channel. Americans tend to forget that it's a big world out there, and the UAVs are legal in just about all of it.
Just some thoughts from me as a pilot...
The FAA has mandated the use of ADS-B (Automatic Dependent Surveillance-Broadcast) for all planes, unfortunately they have set a date of 2020 for implementation. I believe it is already required in Europe. I really wish they would require it earlier. It does require a substantial investment in equipment for both ADS-B In and ADS-B Out, but the increase in safety would be tremendous.
I have an ADS-B In device (Garmin GDL-39) that will display traffic (and weather) on a tablet. However, the way ADS-B was setup by the FAA it will only display traffic if there is an ADS-B Out equipped plane in the area, which then triggers the nearest tower to send out the traffic info for the area. I have gone on long cross country flights and sometimes there are several planes with ADS-B Out around me and sometimes none. In order to be "official" the ADS-B In needs to be connected to a panel mount display, the three planes I fly all have Garmin 430 units, which I believe will accept the data link (with a costly update of course).
So, my feelings are that IF the drone has ADS-B out, and I can see where it is at all times, AND IF a flight plan is filed with the FAA, that would be OK. I almost always file an IFR flight plan when going cross country, but often fly locally with no flight plan. I definitely think certification is necessary, I wouldn't want just anyone creating a flight plan and loading it into the drone and letting it loose without knowing the ramifications and the rules.
The topo-in-the-woods project (which I discussed in a recent post) I am currently working on is for the USACE. Jacksonville District apparently has some type of drone, and will do projects for other districts. This site would be a good candidate, but they declined right away because it is within 5 miles of an airport (FAA rule). Actually, it is right off the departure end of the runway, and I see/hear planes all the time there. Interestingly, there is a RC field adjacent to the site, and I can also hear (and see) RC planes flying.
The first practical applications of photogrammetry were in Europe about 100 years ago for contouring from photos taken from the ground. Of course it can be done, and current digital cameras are relatively easy to calibrate for lens distortion using commercial photogrammetry software. "PhotoModler" and "iWitness" are the two best-selling commercial photogrammetry software packages that do NOT cost an arm and a leg.
Of course, low-level UAV photogrammetry would make it a LOT easier to acquire volumes than purely terrestrial photography.
That's how photogrammetric mapping firms have been doing in on a weekly-to-monthly basis for over 50 years. I used to do it in the 1970s when I ran a mapping company.
I enjoyed meeting you at the convention.
For your ground based photogrammetry, chek out the Trimble V10 Imaging Rover.
It mounts under the R10 GPS head and contains 12 cameras mounted in the circumference.
This enables photogrammetry stereo models in Trimble Business Center 3.0.
The V10 is too new to have had any at theconvention, but it is viewable here:
Trimble Imaging Rover
A couple of months back I went to a demo for the DJI Phantom. RC based photogrammetry for about $1100. That is the Phantom for $700 and a GoPro 3 for $400.
DJI has a new RC being developed called the DJI Vision (built in camera).
I have PhotoModeler software which works well with a calibrated camera.
You can get a demo version for use for a temporary time.
PhotoModeler
Interesting 1998 article showing some
Interesting technology.
FWIW, here is the link to an interesting article describing the use of an UAV for targeting. It covers many issues relevant to this thread.
http://techdigest.jhuapl.edu/TD/td1904/criss.pdf
The technique: Multiple Image Coordinate Extraction (MICE) was being modified for peaceful uses when I encountered the lead author. He proposed a terrestrially-based system where a boom-mounted camera would be used rather than the UAV. The idea was a little ahead of its time given the limited computation capabilities.
HTH,
DMM
Thanks for the information on the V-10.
Bryan! It was great to meet you also!
Do you think the PhotoModeler Software will be able to correct for the distortion of the GoPro ? I Have a GoPro 3 black with all the bells and whistles. Would the wide angle be more advantageous to the data or would it be harder for the software to find angle solutions?
Holler at me sometime when you pass thru the Brady area.