Comment by narwhal on June 22, 2011 at 8:37pm

@Krzysztof Bosak
When I say real time obstacle avoidance I mean if the UAV is flying towards a tree it can gather the models in time to plot a course to evade it

Comment by Clark Hymas on June 22, 2011 at 8:40pm

Precision agriculture - excuse me for going a little off topic but this all ties in with the IR and what it us used for in our farming opperation.

Some of the things we are doing with GPS:  at this time we are planting with 5 tractors/planters in the same field starting at 5 different locations all coming together with the row spacing being within 1 inch acuracy.  At the same time we are also controling seed rates and fertilizer being applied.  When we get done with each 125 acre field the amount of seed applied is within 1.5% and the fertilizer being within 0.7%.

As these crops start to grow and we take the IR photos you can see spots where the seed may have been planted too deep or one row that had a clogged nozzle.  The percentages above were from some of the better fields, the field that are out of spec are the ones you can see variances in.  When planting 8000 acres of spuds, 9500 acres of corn, and 5500 acres of peas it pays to be precise and be able to analize the results.

@ me - Like I said before, I'm not the camera guy and haven't dove into that one much yet, just going by what I was told.

Comment by Krzysztof Bosak on June 23, 2011 at 1:42am

@Clark

For me this is exactly ON topic :-)

However, for detecting one inch nozzle track you need some half inch per pixel... Right?

Comment by Lojze Miklavčič on June 23, 2011 at 2:30am

Well I guess not. Except in a case when rows are also one inch apart. 5-10cm GSD should be enough for this.

Anyway NIR image in grayscale color is more useful in analysis than false coloured one.

Comment by Jonas Beseler on June 23, 2011 at 3:49am

Hey Krzystof,

I have a good idea. I am working in the wind power industry. I thought about the guys that are doing the wind prediction models for new wind farms. They need information for the surrounding area to make their wind prediction models. Here would probably 3D models from the surrounding terrain very useful.

Maybe you can talk to the guys at www.cube-engineering.com they make the software for such calculations.

Hope that helps with your crisis

Best Regards

SirSten

Comment by Lojze Miklavčič on June 23, 2011 at 4:35am

... and I'm working in visualization of those and comparable objects (powerlines,...) using LIDAR data. Investors are regularly under attack from local communities due to visual impact ;-) For smaller projects, your services Krzystof can be quite valuable. The main problem that I have is how to extract from orthos and LIDAR data tree types (conifer, deciduous, bush) and NIR can be here (in right conditions) very, very valuable.

Comment by Krzysztof Bosak on June 23, 2011 at 6:59am

@Jonas, Lojze, excellent ideas. Will be pursuing in that direction. What worries me,having full spectrum is beneficial for having short exposition times (blur is an issue in smal UAV).

I am wondering if subtracting VIS from FS will bring the same NIR quality I would get by using optical filters (that otherwise pass little light). This is an important issue that might affect flight plan and several mechanical details.

Comment by Jonathan Lussier on June 23, 2011 at 8:13am

Chris,

what's the flight speed of your aircraft?  Maybe the blur could be reduced at low speed (30km/hr) and slightly higher altidue.

Comment by Krzysztof Bosak on June 23, 2011 at 8:37am

Airspeed 45-55km/h

Major source of blur is turbulence.

Comment by MarcS on June 23, 2011 at 8:47am

Chris,

Substracting VIS from FS will not give nearly the same quality as an optical filter (tried and compared myself...). Mostly because you got much Vis remainig in the image. The calculation is just not accurate enough (and the efficiency of the sensors in IR too low). You just can´t trick physics :-(

If you want a high sensitive chip in IR you have to go for professional stuff, you know the prices... It´s just not mainstream.

If you want to do flying for remote sensing try to do it when the sun is in the zenith. You get the least shadows and least effects of BRDF´s...

And if a system is as robust as a tractor and usable by a farmer we can talk about precision farming. As long as you need an engineer/pilot to operate and prepare it, it will stay an academic use...

Comment

• ‹ Previous
• 1
• 2
• 3
• 4
• Next ›
• Page