Lately I have been getting a lot of requests to do rooftop surveys and was wondering if it would be efficient/practical to use a drone to capture the data and, if so, what kind of drone should I be looking at buying and what kind of accuracy would I be getting.
?ÿ
?ÿ
At a drone seminar in December, I learned that the different locations determined at the top of a two story building was over 0.3ft out of place in relation to the control on the ground.
With pictures, the focus on one plane "the ground" will not generate good data for another plane "the roof".
For what you are wanting, you would probably need to acquire data in relation to different control planes for data collection and combine them for a finished product.
0.02
?ÿ
It will be used to map the rooftop only and I will have control on the roof, most likely the top corners of the building surveyed in with GPS. Sometimes there is a lot of features on the roof they want mapped.
There are a lot of reasons for surveying a rooftop. Ive done it for everything from cell towers to pools to HVAC retrofits. Use the tool that gets the required result efficiently, just like on the ground. I can see a drone giving you what you need for some applications.
this is for cell tower purposes - mostly finding a usable space for their equipment on the roof.
this is for cell tower purposes - mostly finding a usable space for their equipment on the roof.
I would think it is an excellent application?ÿfor drone mapping. If this roof has lot's of features and changes in elevation, such as a hotel, then mapping this conventionally would be tedious?ÿand prone to missing something.
What type of drone and software would I need to get this done??ÿ
DJI Phantom 4 Pro and an iPad tablet would get you started. Then you can send the photos and your control points to me and I could process it for you.?ÿ
Or you could purchase pix4d for $8,700 additional cost.
You would also need your FAA part 107 certificate.
Have done a few "roof top" surveys in high rises to determine obstructions
to towers.?ÿ And forgive me if I sound condescending.?ÿ?ÿ
Establishing position on the roof only requires you to be on the roof.
Building is /or should be in a somewhat geodetic position as far as area
that it entails??ÿ I like drone approach, however, in this case seems to be a
bit overkill.?ÿ Apologies if I misunderstood.
DB
?ÿ
Here is a link to a detailed Drone Rooftop Survey with Pix4D?ÿand?ÿDJI Phantom 4 Pro.
"65-degree camera angle at 80/80 overlap at circa 15m above the roof level giving an average GSD of 6.2mm"
At a drone seminar in December, I learned that the different locations determined at the top of a two story building was over 0.3ft out of place in relation to the control on the ground.
With pictures, the focus on one plane "the ground" will not generate good data for another plane "the roof".
For what you are wanting, you would probably need to acquire data in relation to different control planes for data collection and combine them for a finished product.
0.02
?ÿ
This is only correct if you are just using "pictures".?ÿ Creating a point cloud from the photography enables accurate levels to be extracted from any elevation within the model.
The only provisos are that for SUA data very thin "point" objects such as lamp and telephone poles MAY be poor, but not always. There is sometimes a problem with overhead conductors as well.
Otherwise roof heights should be accurate relative to the ground control to within around 0.1 ft (depending on your flying height and the camera. Using a Sony A6000 from 320 ft. I would expect no problems in achieving that level of accuracy). You do need to get good coverage: standard overlaps should give you something like 8-10 photos showing each point. Remember to fly at least two lines beyond the outer edge of your area of interest to ensure the full coverage is there. It's easy to get the extra shots while airborne, much more effort to refly to fill in a gap.
Thanks Lee. The detail looks good, even if you zoom way in.?ÿ
Antcrook,
They flew a very?ÿlow Altitude for high-resolution?ÿGSD.
You can almost read the labels on the trash cans.
I flew proposed rooftop cell site and compared the field measurements with the rectified orthometric photo from pix4d and the two surveys compare really well, within 0.1' horizontal.?ÿ?ÿ
https://cloud.pix4d.com/pro/project/243539/map?shareToken=3c1b8acb9b524324811fa0f0509717a8
Very cool images, Lee!?ÿ Is that you in the lower right??ÿ ?ÿI see all the painted control point circles.?ÿ ?ÿ
