I'm seeking to test UAV acquisitioned data, primarily captured by high end digital cameras. I'm an expert in photogrammetry and remote sensing and if you have the capability of acquiring such data, feel free to email. I would be very interested in UAVs for accurate 3D geospatial use.
terrainsolutions@gmail.com
William
Replies
Hi,
I started an aerial mapping activity with a low cost Iris+/Sony RX100 20MP solution and it gave me very good results. My concerns is now about time flight limited to 10 min giving me 10 ha coverage, and I am working on a second quadricopter to reach 30 min.
Honestly it was really a trial platform in the begining but it appears I can use it for production. That is a less than 1500 € solution. Concerning camera, the professional Ebee uses even cheaper Canon 12MP. As Marc said, it is preferable to get more pictures than focusing in high resolution (and more costly camera). What I see as an advantage of quadri vs plane is low speed, meaning more pictures. Of course for larger project, a plane is a must. This is my next planned purchase.
Pascal, thanks for your response,
Between Marc and yourself, as I stated before "I feel like an Infant" in this realm. Maybe soon, I'll be able to reach your level of expertise. Now, camera resolution, to me, would seem important if processing 3D models for metrics through semi-global matching algorithms. I read, somewhere that a 1 to 2cm GSD was possible with a NEX5 camera, but now I understand from Marc that the NEX7 is no longer in production. I endeavor to begin with stockpile volumetric computations. I sign and seal volumetric reports by statute in Florida. Is this a requirement from where you operate? My questions seem to be literally all over the map! No Pun Intended.
Thanks again, I'll try to organize my questions a bit better.
Appreciate your response
William
For further infos, contact me at: info@luftbildfotograf.com
With regards
Marc Kulhavy
Marc,
Thanks for the information. Would a hexacopter be suited for quarries? I'm still in decision mode. I probably will go with a Sony NEX 7. Do you have a screen capture of the camera file? Is it self calibrating or will it require one?. Sorry for all the questions! I'm a certified photogrammetrist just trying to break into the UAS realm.
Thanks again
William
I found out its best to let the software do the lens calibration for every job dynamically rather then doing a fixed precalibration. As temperture, aperture and focal point vary, a precalibration makes only sens in terms of a quicker first alignment of the dataset. Afterwards, a dynamic calibration does a suprisingly better job than the precalibration.
The Nex7 is no longer in production, I would stick to an Alpha 6000 or if the price does not matter, an Alpha 7R, which costs triple the money.
Going to big with resolution, sensor size and stuff like than has one disadvantage. You get more information per shot, but for a desired resolution, you will need less pictures, and that means less different angles and views. But this is what the software needs to work with.
The nice thing with a multicopter is that you can vary the camera angle, so one can adapt to the typical shape of a quarry. If there are no vertical or near vertical edges, an airplane can do the same job.
If you really want to step into mapping with drones, you will need both. I started with multicopters first, but the first bigger jobs were all done with airplanes I added to my hangar quickly. It strongly depends on your personal preferences and the type of jobs you will have to deal with.
With regards
Marc
Marc,
Thanks for your insight. Even though I may be an experienced photogrammetrist, I feel like an infant in the UAV arena. Your comments are making me more the wiser in moving forward. Is there a repository of UAV imagery for metric mapping somewhere that can be used solely for testing? I have a suite of photogrammetric software from planning to ortho, DTM, DEM and DSM production. Around 2004, I had an opportunity to produce orthophotography from a predator drone. The systems were lacking IMU, INS and believe or not GPS data. I had to use my creative license to make it materialize. Again, I appreciate your great insight. I'm forging ahead and will take your advice.
Thanks ,
William
you are welcome. We all are here to learn and share our knowledge. The UAV-newby-feeling is nothing to be ashamed of. Even though I build my own models for over 25 years now, i was sweating like sitting in hell when I made my first small hops with a multirotor system or let my big flying wing take a tour via autopilot across the airfield.
But don't forget one thing it's a tool. Not more, not less. Many people here around have other reasons for flying with an UAV, but doing photogrammetric surveys renders the used aircraft to nothing more than a rig that holds the camera in the desired place and points it in the wanted direction.
Of course, you should know what you do when you let a bundle of money lift off the ground, but still, the real result ist being produced at home in front of a workstation computer or even a render farm in an office.
There are different things to concider when you do this type of jobs. For examle, the very good Alpha 6000 has a 24 Megapixel sensor and used with a 19mm Sigma lens, it produces stunning sharp pictures. But as you tend to fly low, you run into new problems like trees and bushes. The classic photogrammetry with GSDs of 10 to 20cm had no problem with trees and forests. The result showed nothing at all from the tree or a closed plane where the tips of the trees in the forest would have been. But now, with super high resolution and low flight heigts, the same forest looks like a point cloud massacre and if you try to generate a mesh, it is not a nice view for the eye. So depending what you need, you can shoot even more pictures to get a cleaner shape, or you can lower the resolution to get a good approximation of it, but with lesser detail.
Another thing to concider is the level of technical understanding and knowledge you need to have - or at least you should have. Multicopters are technical beasts, 4 to 8 motors, motor controllers, power distribution boards, a compex airframe, a camera gimbal, small space for the components, low flight times. If something fails, the thing usually takes the straight way down and does not stay in one piece after touchdown. An airplane has one motor with controller, a couple of servos and a camera without gimbal pointing straight down, thats it. If the motor fails, you glide down, if a servo fails, you still have some control. And if the landing doesn't go ad well as planned, the foam material the most parts are made of are relativel forgiving.
If you want to test some imagery with your software, contact me via pn, I can send you at least a part of some projects I did for testing.
With regards
Marc
Thanks, I appreciate the few who understand what I'm seeking.
Hello,
Have a look at the UAS offerings from Falcon UAV (us), Pteryx UAV (eu) and / or Event 38 UAV. These 3 platforms are very stable, they have long flight duration and great mapping features. We've seen and processed many data sets from these platforms with ease. I would also recommend building a small multi rotor mapping platform based off of the DIYDRONES APM if you are interested in mapping smaller areas / sites. Most systems would start with a 12 mega pixel image sensor like the Canon Powershot, these are great and work well (easy to control with CHDK)
Of course it never hurts to use an imagery sensor like the NEX7 as well.
-
1
-
2
of 2 Next