Bandwagon's Drones Thread

Picasso_sl said:

I thought you had some special hardware for this. You're just taking pictures and having the drones gps provide gps and there's software that'll generate georeferenced dems based on that picture and the drones gps?

Fucking amazing if that's the case

Click to expand...

The special hardware is for Ag. The low-end is converted cameras that have the factory filters removed and replaced with filters that allow NIR or red edge light through. This is what's used in NDVI. The high end cameras are radiometrically calibrated and have narrowband filters with high spectral resolution, typically one band per sensor.

For survey work, people are using high MP cameras if they're sticking to photogrammetry, and putting a lot of effort into getting "perfect" GPS coordinates inserted into the picture exif data, pulled from the flight controller. Personally, I don't see any way around using ground control points with high accuracy GPS for survey work, even if the drone itself is using RTK gps with a basestation (like the RTK eBee).

Really, the drone doesn't matter at all. What I usually tell people is that "You can hang a camera off a balloon or kite and get the same results, but the process is more difficult". What is important is the sensor and the process....the drone is just a "quality of life" variable for the person that is actually gathering the data.

The software does a great job of generalizing DEMs based off of known camera parameters extracted from the exif data, but real accuracy still requires GCPs. If we're talking about NDVI/Multispectral sensing,real, accuracteresults also require a narrow band filter, RAW outputs, an irradience sensor, ground calibration targets, and data processing that doesn't distort pixel values to create an output that is pleasing to the eye. Also, you'll want high accuracy GPS if you're trying to create management zones for variable rate applications.

If you're just trying to make a pretty map, you can do that with a $1k drone and a $40 payment to a data processor.

edit: and for most of the stuff in the past, I use the S100's onboard GPS logger to georeference the outputs. There are ways to geotag your pictures if the camera doesn't have built in gps (like missionplanner's geotag tool), but all of those options are way too inaccurate for survey work. The outputs that I post that have good geotagging are either edited with qGIS or Pix4d using WGS84 GPS coordinates from google earth, and are equally as useless for survey work....but they look better when displayed on a map.

The biggest advantages to lidar is vegetation penetration and accuracy without the need for ground control points. Without GCPs, I don't think you'll even come close to lidar. One of the examples on my link I posted earlier is from a phantom. Definitely the worst quality of the bunch.

Two fun things today- I just got an email from Lockheed Martin's unmanned division in Utah about getting a "loaner drone" this summer. I have no idea how realistic this is and if we're big enough for them to actually do it, but it's pretty exciting to see an email from Lockheed in my inbox.

Also, our new guy wants to start gathering info and pricing on "the big boy drones", so I'm swinging into Aerovel in about an hour to look at the flexrotor, then driving to Insitu to look at the Scaneagle afterwards.

That's the level I've wanted to operate at since first starting down this path, but it's a little scary seeing it as a possibility this early. The other two guys have the aviation experience to operate at this level and the new guy might have the business experience, but we're talking about jumping from a company with around $60,000 in assets to over $1mil.
I'm all for the "fake it till you make it" mentality, but there are a few key areas I would like to see some more experience in.
Scary stuff.

Here is the Oppenheimer report on the current drone industry and forecasts, if anyone is interested.
https://opco2.bluematrix.com/docs/pd...d922fae1be.pdf

I'll give an update tomorrow.
Here's something I was working on after getting back from Portland.

That last project got a fish on the hook.

edit- changed video to a quick combination of step 1 & 3 of processing.

What are "cut volume" and "fill volume"?

Also, wtf does a quarry want? Depletion estimates?

Here is what i was talking about the other day:
http://www.reuters.com/article/us-al...-idUSKCN0VJ00H

I haven't calculated them yet. I just had about 20 minutes this morning before I went in to work.
These guys occasionally calculate volume, but it's not a high value product. They use a survey company for all of their permitted paving projects, which is a lot. This specific one was to expand the 13.9a quarry to 30a.
I'll check the link when I get back to the office.

LachiusTZ_sl said:

Here is what i was talking about the other day:
http://www.reuters.com/article/us-al...-idUSKCN0VJ00H

Click to expand...

I doubt that will have any effect on the next few years for UAS....but hopefully that will mean that I can fly to Portland while taking a nap in my personal Hexacopter by 2030.


and sorry, i missed the question about cut vs fill.

See below, but basically - One measures +volume, the other measure -volume.

Thats about what I assumed cut and fill were.

Seems useful if you build a lot of lakes etc.

The watershed council is interested, they would like to use us to estimate proposed resevoir capacity. I want to have 2 or 3 in depth projects where we compare results with a traditional survey crew, so we have some verification on accuracy.

That's twice that I've 404'd trying to type a long update. I guess that's a sign to keep it short.

Anyways-
Lockheed - No word back after the initial few contacts. It's a long shot and I don't care too much (Indago is the UAS we would possibly get, nothing too fancy). Name dropping is the most appealing part.
Insitu - they had a private tour going on, didn't even get in the door.
Aerovel - The guy I needed to see was gone, but texted me before I even left the parking lot. we chatted for about an hour. Summary was- for CONUS operations, it's too early to plan flight ops like what we'd want a $$$ drone for. We need FAA authorization for BVLOS and night ops.


Also, my two biggest tips for anyone thinking about doing something similiar to what I've been posting (since you fuckers only PM instead of posting ):

1. Be honest about the limitations of your experience, knowledge and drone/sensor setup. A lot of potential customers are skeptical, but almost all of them are waaay more receptive when they realize that you're not a salesman.
2. When you try to learn more about a specific topic/task/request, look OUTSIDE of the drone industry. I had no idea that I didn't really know shit about drones & agriculture until I started reading outside white papers. I learned a lot more that way, and it was not biased sales pitches. The same is holding true for learning about survey work. At the very least, learn enough to know what you don't know.

I kind of hate the real estate stuff- I can't really do the post-production stuff, since I'm colorblind.

I would think fill could be any number of things.

Before / after, used with tonnage to confirm the amount quarried?
Or for depletion estimates for the coming year? Or for planning on how to quarry based on the estimated rate of extraction? Or to look at how much will be required to refilled the quarry after X tonnage extracted?

They mainly wanted the map to plan on expanding the quarry, so the volume estimates are just something I'm going to be comparing with their survey crew. I haven't even figured them out yet, and I'll have to mark every pile and piece of machinery to subtract it out.
Hopefully I'll get to work on it tomorrow....been working on trying to figure out how to post-process photos and videos for people that can see colors.

One longer continuous shot of the same house. We're mainly practicing shots and editing (or trying to), but I also wanted to save a bunch of flight paths that look good, and then actually shoot them during the growing season.

Was supposed to be a new feature, but I can't figure out where the hell to save them.