"This is for everyone." In need of sample data: raw gps/attitude APM2 log file and downward-facing images or video (preferably a GoPro HD Hero 2)

The title says it all. I'm working on an open source mapping and analysis software and I need two different sample data:

The raw log file from an Ardupilot Mega 2 board and downward facing aerial images or video from a flight. If someone has these, you will be greatly appreciated and remembered as a contributor to this software.

Tags: data, sample

Views: 1086

Reply to This

Replies to This Discussion

how do you want the flight, Stabalize, loiter, wp guidance?   Any particular camera angle?  I can get something up tonight using a swann freestyle at either 720p/60fps or 1080p/30fps

Hi Carl,

Thank you so much!

Here is how I want the flight:

  • Camera facing straight down to the ground.
  • Camera fixed to the body (no hardware stabilization unless it's so good that all turbulence will not result in any roll or pitch changes--that is roll and pitch should remain almost perfectly 0 if you have stabilization) If there is no stabilization my software will correct images that are taken at an angle. Thus, anything in between, my software won't know what to do.
  • Top of camera pointing towards front of board/vehicle -- so when yaw is 0, the camera takes an image whose top points North. If not then I'd need the degrees offset.
  • Preferably 720p / 60 fps (higher fps, clearer image)
    • I assume you don't have a GoPro HD Hero2 if you made a typo and meant 720p/48fps which is the highest fps for 720p mode on the Hero 2. If you do have a HD Hero 2, then everything is fine and use 720p/48fps.
    • If you don't have a GoPro HD Hero2 things are more complicated for you and I understand if you don't want to take the following steps. Can you tell me the specifications of the camera? Is there a fish eye effect? What is the horizontal and vertical field of view?
      • If it has any type of lens distortion, I can tell you some steps to take to figure out the distortion parameters via calibration script I have.
      • If you can, can you figure out either the sensor size and focal length or manually calculate the field of view?
  • Fly in lawnmower fashion at a steady altitude
Video
Video Format
Video Format MPEG4 Codec, MOV File Format
Resolution
1080p (1920 X 1080) 30fps
720p (1280 X 720) 60fps
Photo Format
JPEG
Photo Resolution
Photo Resolution 8MP / 5MP / 3MP
Capture Mode
Single Shot, Continuous every 2, 3, 5, 10, 20, 30, 60 secs, self timer every 0, 5, 10 secs
Optics
135 Degree angle view at 1080p resolution
170 Degree angle view at 720p resolution
Audio
Microphone
Yes - Built in
Audio Format
Stereo 48kHZ, AAC compression
Recording
Storage Type
MicroSD Card (not included)
Storage Capacity
Up to 32GB
General
Record Time
Up to 2.5 Hours per charge
Battery Type
Rechargeable Lithium-ion (1000mAh)
Charge Method
USB (computer or charger)
Dimensions
2.36" x 2" x 1.6" / 60mm x 51mm x 42mm
Weight
2.9oz / 82g (without screen)
3.8oz / 108g (with screen)
Storage Temp
-4 degrees F ~ 140 degrees F
-20 degrees C ~ 60 degrees C
Working Temp
-4 degrees F ~ 122 degrees F
-20 degrees C ~ 50 degrees C (When inside waterproof housing)

I picked it up from a wootoff for about $100, My mount is just a regular static mount like you would stick on a helmet.  I can pivot the camera so that its lens is level facing down with no mods so ill shoot some data tonight and provide you with the dropbox link to the zip tomorrow. 

This is what I shot over the weekend with the camera at a slight down angle.  Video was reduced to 640x480 for youtube. 

http://youtu.be/_GobGHh3HPc

Can you take two super short videos (one 720p/60fps and one 1080p/30fps) of your screen displaying this calibration chessboard in fullscreen (press F11 key) mode? http://1.bp.blogspot.com/_0zx-4pM8HnA/TSltkqjXscI/AAAAAAAAAI0/45MJo...

And then upload them somewhere for me to download.

If you can, can you measure the physical width and height of one of a 2x2 block of squares on your screen, and the distance away (at least a couple feet) your camera lens is from the screen? Optimally, you're holding your camera lens surface parallel to the screen and at the center of the normal axis of the screen.

Also: Can you disable the onscreen date/time display?

Ill do it tonight when I get home.  I have a 23" inspiron one and a tripod so everything will be nice and level.  The camera has a laser designator for aiming approximate center of frame +/-1 inch 

Ok, wow, thank you. I really appreciate this. I'll let you know of the test results once I process the video and telemetry data when I get it. My code is ready to go--just needed someone like you to offer sample data.

Ok here are the constants. Sorry i didnt get it done yesterday,  I used a hp laptop with a 15.4" screen (391.16mm) for the display of the checkerboard.  The width of 2 squares is 53.5mm, the camera was 250mm away from the screen with the lens as parallel to the display. this is as stable as I could get it by hand.  the laser dot on the upper left is approximate center with the dot on the lower right the reflected beam from the display.  

The download link for my video clips is below.

http://dl.dropbox.com/u/86596/sample.zip

Let me know what resolution you want and Ill get some video shot over lunch 1pm central.  If not It will have to wait till tomorrow.  

Thanks for doing this. I think the 1080p 30 fps mode might be better for the flight. That seems like one that has the least natural distortion. Make sure the on screen display is disabled during the flight, camera faces down and the view of the ground during flight is unobstructed by parts of the vehicle.

I should have something up by Saturday evening at the latest.  Seems that we are finally getting some of the overdue rain that we have been needing.  There is rain blowing in currently, and it is supposed to subside by Friday. 

Can you do the calibration video and measurements for the 1080p mode more accurately? This is actually quite crucial to good accurate results.

Here's what I'd do:

  • disable on-screen display
  • to reduce glare print out image onto horizontally oriented paper
  • tape paper to flat wall
  • take video of paper in 1080p mode at a certain measured distance away making sure entire board is visible in the video but also filling up the video as much as possible
  • make the same measurements again

I'm sorry about asking you to do this again--however if done accurately, the results will be good.

Ok, I have done the above steps again.  The width is 51mm for 2 squares.  The camera is standing on a stack of CD jewlcases and the lens is 250mm from the paper

 

http://db.tt/9PIodVxs

OK, thanks, I think I've got the calibration right.

Before:

After:  I'll just need to crop the edges down I guess. Unless you can do the same calibraiton but on a larger piece of paper so that the entire image is chessboard squares. Getting too close to the paper is not a good idea because the focus of the camera will be practically infinite during flight--much greater than when focusing on this board.

Anyhow, though important for accuracy, currently this is not the biggest issue at hand. Thanks for doing this calibration and I can't wait until I use your data for my first test of my program!!

RSS

Social Networking

Contests

Season Two of the Trust Time Trial (T3) Contest has now begun. The fourth round is an accuracy round for multicopters, which requires contestants to fly a cube. The deadline is April 14th.

A list of all T3 contests is here

Groups

Advertisement

© 2013   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service