Thread Tools
May 10, 2008, 10:58 AM
Registered User
dmgoedde's Avatar
Thread OP
Sim - did you see the 3D photos in this thread? They are some of the most recent images posted. These are stills from video, seperated by a few seconds. Low quality for sure, but it demonstrates what is possible.

So to do what you mention, then I envision taking pairs of photos (specified distance seperation, say 25 meters) and also being able to specify distance between adjacent pairs (say 400 meters).

Dean
Sign up now
to remove ads between posts
May 10, 2008, 01:03 PM
Inherent Tinkerer
Dan, <---- Oops! Should have been DEAN! :-P (That's what happens when you have to get up at 0500 to go teach someone how to run a business jet in the simulator!)

Your 400m solution is exactly what's needed to do aerial mapping. Being able to store the aircraft's location and attitude, as you've previously mentioned, will allow us to easily create a mosaic using software that's already available. The biggest challenge is determining the proper focal length of the lens to keep from getting any distortion (fish eye). The distance between shots will, of course, depend on aircraft speed, altitude, and the camera lens.

Concerning 3D images, I think it's possible to set up two cameras side-by-side (fore and aft?) and adjust for parallax correction to be able to get a 3D shot. Have to do some experimenting to test it out.

Jimmy
Last edited by jtprouty; May 10, 2008 at 03:56 PM.
May 10, 2008, 01:56 PM
Professor of Wood
kd7ost's Avatar
Agreed,

And I think Jimmy meant Dean.

The potential big problem in doing 3D images from a single camera is in the platform. With pitch and roll it will be hard to frame the second shot like the first shot unless there is something employed to keep the Camera pointing nadir. A camera pan and tilt would likely solve most of it if used with something like the Unav Pico Tilt unit.

But as Jimmy pointed out two camera's that have been adjusted for parallax angles it would be easy to repeatedly take good 3D images as long as you get both camera's to fire at the same time.

But the concept of having a user adjustable trigger distance activation is still a great idea. Much work can be done without having to use 3D images as far as mapping. But there are lots of users out there that would be happy to have 3D images.

Dan
May 10, 2008, 04:06 PM
Registered User
Quote:
Originally Posted by dmgoedde
Sim - did you see the 3D photos in this thread? They are some of the most recent images posted. These are stills from video, seperated by a few seconds. Low quality for sure, but it demonstrates what is possible.

So to do what you mention, then I envision taking pairs of photos (specified distance seperation, say 25 meters) and also being able to specify distance between adjacent pairs (say 400 meters).

Dean
hi
yes I did see them but cannot get my eyes to relax enough to do the over lap.
(specified distance seperation, say 25 meters) and also being able to specify distance between adjacent pairs (say 400 meters).
This would be ok too but we could also just set a trigger to start taking photos at a set way point every second or so for as long as we want. then we could just use any 2 that were taken in sequence or use all to make a mosaic.
With Atto flying and if wind is calm the plane will stay level during the run. If not then gimbals could be made to keep camera level But If you look back at my stitched photo, Most were not level as the plane was in a turn. The softwares corrected most of that. Again we did not go out to make that photo.
Thanks
Sim
May 10, 2008, 08:10 PM
Registered User
dmgoedde's Avatar
Thread OP
I'll have at least 3 trigger options:

1) At WP
2) time interval repeat
3) distance interval repeat

and then in addition to those there is:
a) force flat roll + pitch (at expense of navigation temporarily)
b) Flag into the LOG file to indicate photo shot

I think that covers it pretty well, definately enough for an initial release.
May 10, 2008, 11:21 PM
Registered User

That's what I am looking for


Quote:
Originally Posted by dmgoedde
I'll have at least 3 trigger options:

1) At WP
2) time interval repeat
3) distance interval repeat

and then in addition to those there is:
a) force flat roll + pitch (at expense of navigation temporarily)
b) Flag into the LOG file to indicate photo shot

I think that covers it pretty well, definately enough for an initial release.
Dean,

You are right on target with what I need. I anxiously await my AttoPilot.

Dennis
May 11, 2008, 12:01 AM
Registered User
Hi
yes sounds great
May 12, 2008, 01:02 AM
Registered User
dmgoedde's Avatar
Thread OP
Quote:
Originally Posted by jetjock51
Dean,

You are right on target with what I need. I anxiously await my AttoPilot.

Dennis
It is coming... hardware and coding is moving faster than ever, though I'm not splurging on details here like I used to do.
May 12, 2008, 01:28 AM
Registered User
Quote:
Originally Posted by dmgoedde
It is coming... hardware and coding is moving faster than ever, though I'm not splurging on details here like I used to do.
It's okay Dean, just keep working. You are only allowed 4 hours sleep per day.

Dennis
May 13, 2008, 03:43 PM
Registered User


Need input!

Actually, output

I was curious if, aside from logging, there was some output terminal coming from the Atto. Say for instance, you wanted to have your camera level automatically based on the tilt info Atto was getting from the sensor, would this be possible, maybe as an add-on module.

Thoughts?

Chip-
May 13, 2008, 05:35 PM
Registered User
dmgoedde's Avatar
Thread OP
Quote:
Originally Posted by carguy84
[img]
I was curious if, aside from logging, there was some output terminal coming from the Atto. Say for instance, you wanted to have your camera level automatically based on the tilt info Atto was getting from the sensor, would this be possible, maybe as an add-on module.

Thoughts?

Chip-
Definately - an add-on could intercept and parse the serial data stream on its way to the input line of the modem. But why not just have Atto drive the pan and tilt servos? All you'd have to calibrate is a center position (level) and scaling factors to get the angle right. There are easy ways to teach this on the ground.
May 19, 2008, 04:02 AM
Registered User
fuhan2000's Avatar

attopilot feature creep?


Guys,

I am getting a bit worried about feature-creep.

Let try to focus on the core purpose of attopilot:

To get from one point to another point with smooth flight controls on roll, pitch, yaw axes (just like what we saw on icebear's video on post#696). All the other stuffs like auto-triggering of camera, integration with OSD, auto-levelling of camera are secondary.

The great thing about a forum is that we can have many people exchanging ideas. I like that too. But therein lays a problem. It may cause us to lose focus.

U can have
A. attopilot by itself (with no OSD) - great for navigating between user-defined waypoints and maybe trigger a camera (or drop paratroopers) when say u are x meters away from a target. Also have a "return to home" feature.
B. attopilot with artificial horizon on OSD - great for a FPV pilot that fly over featureless terrain . That artificial horizon will allow him to fly long distance over featureless terrain (like the sea, above cloud) The pilot lets the plane fly by itself while he activates the pan & tilt servos to enjoy the view.
C. attopilot for AP guys - These guys want to have programming features that allow them to create mosaic images of earth surfaces ( a l.a google earth). This assume that their camera can capture image at regular interval (say between waypoint 5 & 7) when trigger by attopilot.

As for me, if Dean announced that Option A is available for sale, i will order one. Option B Artificial Horizon is great but not required. I am an Electric flyer and I doubt that 20 minutes of flying will allow me to fly far anyway. Maybe i will change my mind if i fly glow.

As for Option C, creating mosaic image is seriously cool. Maybe i will pick that later. But with the plane "drifting" on all three axes , it will be difficult to frame the 2nd shot exactly like the 1st. Maybe software is smart enough to correct for these. I am no 3D mosaic expert. Also if u intend to shoot continuously between waypoint 5 & 7, u need a camera with serious processing prowess especially if waypoint 5 & 7 are a long way off. Even "in P-Mode using One-Shot AF the Canon EOS 400Dís buffer memory filled up after 9 seconds" (taken from http://www.kate-harris.co.uk/continu...n-eos-400d-10). How many AP guys fly with a DSLR strapped to their plane? They must be in a minority. The majority fly with compacts equipped with modest processing power.

The bottom line I am trying to stress is let focus on a release date for Option A. For me i can wait. No rush for me. Currently I am waiting for new kit, motor, esc, VideoTx/Rx to arrive.

Option B & C can be pushed back to a later date.

Just a thought,
fuhan2000
May 19, 2008, 11:11 AM
Registered User

I know waiting for this new tech is tough!


Quote:
Originally Posted by fuhan2000
Guys,

I am getting a bit worried about feature-creep.
fuhan,

I brought up the same thing a while back. Dean assured me he has an unannounced production schedule and will not allow these musings to delay it. So then I participated in more musings. I'm just trying to pass the time productively while waiting for the first release! He said we'll get them with time to fly them before it gets cold again. I hope he makes it!

Paul T.
May 19, 2008, 02:11 PM
Down under under
Too late for us down south, NZ to be exact
Never mind, I am sure we can make a water proof plane.
May 19, 2008, 07:38 PM
Registered User
dmgoedde's Avatar
Thread OP
While I have multiple "future" things working on a high level, the basic core is fully intact.

Rest assured that I am working through iterations of a final design. I have distributor and manufacturing capability defined. Legal stuff is moving. All the pieces of puzzle are there, and the schedule is on track.

The camera trigger is the whole point of having a UAV in my mind. Otherwise all you have is a plane flying itself around. Probably 90% of the people interested in Atto want it for AP, and the other 10% as a fancy fail-safe or for FPV so they can focus on the video camera and let Atto do the flying.

The camera trigger capability is a trivial implementation, and will not impact the release date.

There are far more advanced capabilities and hardware sufficiently developed, but I have to save some secrets for the release! I beleive not everyone will want all the features, so there will be various price point versions released.


Quick Reply
Message:

Thread Tools