A drone just happened to show up at our office today. Camera, Lidar Sensor, propellers, blah blah blah.
My coworker has done a bit more digging than I have in regards to postprocessing software but as of now we are undecided. We will be signing up for a few free trials to see what workflow suits us the best. Pix4d and some TBC package are two that we will be trying out shortly.
My biggest question is what are some workflow methods to get your drone data into a printed 2D map. My coworker mentioned that there are some automated tools in some software to identify objects such as curbs, pavement, buildings, etc and automatically create representative 3D linework for those objects. Intriguing - but I have yet to find any good videos or websites clearly demonstrating that capability.
As of yet I can not really find any clear examples of turning any drone data into a 2D map. I've seen amazing 3D models and point clouds but no results translated to paper. Are not many users needing to make paper maps from their drones (just sending out 3D models?) or am I just terrible at searching for this?
Is it practical to create linework by snapping to a point cloud?
Yes, TBC will do automatic feature extraction.
There are some videos for this on youtube
You might like to start with
and
Hope this helps
We use Pix4d and Virtual surveyor to either create the actual 3d lines or points and/or surfaces in xml and then export them to civil3d for production.
We use TopoDOT for the bulk of extraction.
I'd second a Virtual Surveyor to DXF / XML for a great option for general surveying. Only downside is it doesn't have extensive automatic extraction options and that it runs on a DSM instead of a point cloud. It does make the feature extraction relatively easy but you can't click to a point on a cloud to make a vertical wall face, for example.
I don't have experience with Pix4D Survey, but it does have some automatic feature extraction options and it runs on a point cloud base instead of a DSM.
Outside of those, you're generally looking at larger software packages like TopoDOT, Cyclone, etc. to do extractions. They're pricey, but they're also tailored for larger projects. Consider how large your projects will be when looking at packages. We landed on Virtual Surveyor because the price was right, we usually do single intersection projects, largest ones being maybe 1/2 mile stretches.
The real hard part is the QA/QC and being able to express that in your deliverables.
Another option is to send the data to Aerotas and let them do it. It's not expensive as they use your replacement, I mean AI, to do it. You'll likely be shocked at the time and tediousness of processing LiDAR data.
I will recommend that you still have a crew locate any critical and vertical objects such as buildings, fences, curbs, retaining walls, sewers etc.
I will recommend that you still have a crew locate any critical and vertical objects such as buildings, fences, curbs, retaining walls, sewers etc.
Depending on the system used and project scope, we'll have crews locate all, some, or only a few of the critical features.
If you're using solely SfM/photogrammetry to develop the cloud from aerial photographs at 400' AGL, I would recommend it.
If you've got a high-end MMS running at 2MHz collection with photos, you're likely OK with digital extraction and spot checks. Still helps to get boots on the ground to confirm the provenance of certain features.