Announcing T-3, Round 3: The Reliability Round

It's time for the third round of the third round of the second season of the Trust Time Trial (T3) competion! Now that the weather is improving, there are no excuses for not participating. Valuable prizes await!


This the Reliability Round. Our Judge, Gary Mortimer, describes the rules below:


--------------------------------------   RULES   ------------------------

Point of order first: All T3 entries should be flown away from built up areas. Should you have any fly aways or incidents its best to have them miles from anyone else.


The aim of this round is to show reliability.

A very simple task: autonomously fly to a point that you have chosen that is at least 300 metres from your takeoff point (100m for copters). At an altitude of 75 meters (25m for copters) autonomously take a single photo of your point, then return to launch, land (manually or autonomously) and do it all again. Repeat this for a total of five flights.


Your entry should show your take off and landing times for each flight as well as KMLs showing the slight path. As always, post your entries in the comments below.

Each individual image will be compared and the person with the most shots with the chosen point in the centre of the image wins.

Honesty is everything here, don't discover that by chance a point you didn't select is in the centre of each shot and then tell us that's what you were after all along.

But.....

If you can make something interesting with your shots extra points will be forthcoming

You might make a map using MapKnitter Have a look around here to find out what its all about hereDo something with Photosynth, ICE or any of the other similar mapping products springing up. Wouldn't it be nice to have a DIYD generated layer of opensource map images out there!

So in summary one day, five flights, five photos. Extra points for creative use of mapping software/services.

What you do with any other images is up to you. Impress us.


Why the pub challenges in the image above? Well I was looking at a dart board when I thought of it. 

As an aside the area in the picture will be underwater next year, a dam wall is being built on the right. The big circles are irrigation centre pivots just the sort of thing a farmer might be interested at looking at from a UAS

-----------------------------------------------------------

There will be three winners in each category (planes and copters), with the #1s getting a $100 gift certificate at the DIY Drones store and #2 and #3 getting $25 each.

All entrants who successfully compete the course with a copter, regardless of place in the judging, will receive a flying robot merit badge!

Deadline is Sunday, May 13th, at midnight PST. 

Views: 9611


T3
Comment by Krzysztof Bosak on April 13, 2012 at 11:34am

@ChrisAnderson Obviously you haven't shot more than a few dozen photos like that and you are relying on reviews of ppl that never took camera in the air.


3D Robotics
Comment by Chris Anderson on April 13, 2012 at 11:43am

Kryz: I've shot thousands of shots like that over five years, and done the mosaic as well. Here's one of mine from as far back as 2007.  I may not be as expert as you, but there's no need to be insulting, especially when you're wrong.

Comment by Rob_Lefebvre on April 13, 2012 at 12:01pm

For it's day, the S90 was actually a very decent camera, and cost over $400.  Not something that everybody will feel comfortable putting in a model airplane.

However, Moore's law and all that, let's look at a bottom-of-the-line Canon from 2012:

http://www.dpreview.com/products/canon/compacts/canon_a2400is

0.8 fps, and about $140.  I expect the picture quality probably still won't be as good as that S90, but the fps is there, in theory at least. Megapixels are relatively meaningless, the lens is the limiting factor, and cheap cameras like this have crappy lenses.

I do question the real-world usability of these claimed FPS, I'm sure they may be slightly misleading.  I have a Sony WX10, and the specs claim a staggering 10fps, but that is only for a ~10 frame burst.  I'm not sure what the truly continuous rate is.


T3
Comment by Krzysztof Bosak on April 13, 2012 at 12:02pm

Sorry, you are right I am wrong. You can take any compact camera and get shooting rates from photo magazine for trusted. 

Comment by Rob_Lefebvre on April 13, 2012 at 12:14pm

Why are we even arguing about this?  It's not at the core of the mission.


T3
Comment by Krzysztof Bosak on April 13, 2012 at 1:13pm

@Robert Lefebvre "I'm not sure what the truly continuous rate is."

Nobody is because nobody is measuring this and camera tests are just copy-paste commercials by clueless humanists.

Having very high frame rate is the key for assuring sufficient image overlap for map stitching at low altitudes.

Most compact cameras fail below certain...NVM.  already said.


Moderator
Comment by Gary Mortimer on April 13, 2012 at 1:18pm

Blimey I noticed lots of posts on this thread so I thought we must have some entries, little did I know!

I was expecting people to take a single shot really and from just under 400'

Low passes will not be permitted as a restart. Thats just too easy!

I will only accept evidence of five take offs and landings for entries.

This is about reliability and the use of images to make something impressive. It might end up being that least accurate images are turned into the best result and that would have to score highly.

This must be simple exercise for all the photo mapping companies out there. Stand up you lot!

I would suggest that the actual flying ought to be easier than the image production.

I used to use Optios but too many of them went the way of the circular filing cabinet because a good knock causes the lense to jam, it's Canons all the way now for me. Either using CHDK for simplicity or a servo when I want a particular point.

Comment by brakar on April 13, 2012 at 1:47pm

I just looked up the specs on the camera I am planing to use (Canon Ixus 120IS), it is supposed to have a shooting rate 0,9fps in "continous mode".

My experience is however that shooting faster then about 0.5 fps can be problematic. I don't know exactly why, maybe partly due to wind as allready mentioned, trigger method (even if CHDK should be pretty fast), not shooting in "continous mode"?, varying inaccuracy in gps possition, the autopilot, sub-optimal memorycard, camera settings, etc. (But I Can tell it is Very annoying comming home and finding maybe 10 out of 200 images are missing).

But no need to make a fuss out of this, the photgrammetry part is after all not the goal of this round, (even if it is for me).

Comment by János Mészáros on April 13, 2012 at 2:25pm

Guys,

Just my two cents: I'm using Nikon L14 camera (no CHDK, nor manual settings for shutter speed) but it produced very sharp photos three weeks ago during a real field survey. During the surveys, my average flying height was 70 m and my trigger repeat period was ~1800 millisec, depends on ground resolution and overlap. Oh, and I'm using a simple servo finger and the MatrixPilot Trigger action feature to trigger my camera.

According to my experience, I really don't recommend continous mode because the camera focuses only before the first photo and if the relative flying height varies during the flight, it can cause blurring images.


Moderator
Comment by Mark Harrison on April 13, 2012 at 3:29pm

I'm a bit confused.

Does putting the camera into "one shot per second" mode qualify for this trial?  Or does it need to be "take one shot when over waypoint"?

I had thought it was the latter, but am confused regarding the stitching comments, which seem to imply periodic shots.

Many TIA for the clarification!

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2017   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service