Shop OBEX P1 Docs P2 Docs Learn Events
X,Y Camera Tracking System or other words Pan&Tilt Tracking System ... — Parallax Forums

X,Y Camera Tracking System or other words Pan&Tilt Tracking System ...

wperkowperko Posts: 66
edited 2006-06-02 04:31 in General Discussion
cool.gif Hi,

I'm new to the world of STAMPs ... and I have a couple of projects I want to create and am here to poll for help ...

Primarily I want to build a system that controls a Pan & Tilt dual servo setup for a small platform holding as many as three video cameras to track moving objects in the sky.

I've been thinking either an optical tracker or maybe an aural tracking system ... or maybe both for redundancy ???

Maybe the system could track an ultra-high pitched whistle inside the moving object (R/C airplane or R/C car etc...) or a set of specific frequency LEDs all around the object?

So, anyway, I'm open to ideas ...

I'm very new to the STAMP system ... but do know BASIC programming reasonably well ...

I've started with the "Basic Stamp 2 USB Discovery Kit" ... so far I have 4-servos under very simple control to amuse my pet Cockatiels ... basically 4-hands waving at different intervals at the birds ...


So, if anybody can give me more clues on my journey I'd appreciate it ... smilewinkgrin.gif


Comments

  • Tom WalkerTom Walker Posts: 509
    edited 2006-04-26 17:51
    I think that the "front-end" of this is going to be your problem. Somehow, you will have to devise a sensor (or most likely, an array of sensors) that can give you meaningful information about the location of a distant object. Are you designing this "distant object"? More information will be required.

    Once you can get good information on the object's location (including the time domain), working servos is a relatively easy task (assuming no great speed is necessary).

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Truly Understand the Fundamentals and the Path will be so much easier...
  • wperkowperko Posts: 66
    edited 2006-04-26 18:42


    cool.gif Hi,

    Okay, maybe I should make it a little clearer than like I had said in the first post ...


    I want to make a Pan&Tilt platform that can hold as many as 3-camcorders to track in an x,y positioning of things like R/C aircraft in flight, R/C cars or other moving objects within the x,y view of the platform ...


    I need help to learn about different ways to track objects in motion ...

    What kind of sensors can see a moving object? or Do I need to install some kind of transmitter into the moving object so a sensor can track it ???


    I am new to STAMPs but hopefully can program the PBASIC ...


    But first I need help with the hardware that might solve the problem ...



  • Beau SchwabeBeau Schwabe Posts: 6,545
    edited 2006-04-26 20:39
    This sort of reminds me of when each of my daughters were born. ( .... But wait please let me explain!)

    In both cases, in the delivery room there was a gimbaled spot light hanging from the ceiling that the
    doctor/nurse had control over. To position the spot light, all you needed was a flashing strobe light for a
    second or so and the spot light would re-position itself to the source of the strobe.

    Now in your case the strobe light sounds like it could be placed on an RC car or plane.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • wperkowperko Posts: 66
    edited 2006-04-26 21:26

    cool.gif Hi,

    Well, that's sort'a what I was guessing at in my first post ... putting some sort of light or aural trasmitter(s) on the plane or car or any other moving object ... but I don't know how to track a light ...

    When I was in the Navy they had Sidwinder missiles on the jets ... the Sidewinder would follow a heat signature ... the heat signature could be a cup of coffee, a person, a cigarette whatever was the hottest thing in the Sidewinders view it followed and moved servos controlling the flight fins on the side of the rocket ... hence I need help with the hardware that can actually perform that function ...

    I've already mentioned the "Light Source" ... but what kind of detector can tell location and direction?

    Perhaps a large array of detectors ... but that doesn't seem like a viable solution ... I want to bring this system indoors as well as outdoors ... I want to have my camera platform follow the aircraft flight in a 3,000' long x 500' high x,y area outside ... yet still easily setup to follow a 1meter square area indoors too ...


    I realize I will have to start the platform at other than level positions ... even at angles to the x,y axis ...

    I was thinking that maybe "Motion Detection" video camera software could maybe help ... a set of bright lights on an R/C airplane would provide the high-contrast motion ... perhaps the camera could even be adjusted for a specified frequency of light source ... and the software that detects pixal changes in the camera would give position and direction informaiton ... but I don't know what software can actually do this ...

    If anybody knows please post here ...

    I found this particular web site a few years ago ... I've never gotten any response to my questions from this guy who works at NASA ... even though I thought NASA developed public domain information:
    http://www.gvarros.com/meteor.html

    Look at that tracking system ... it looks affordable to build yet they just keep the hardware & software a secret!




  • SciTech02SciTech02 Posts: 154
    edited 2006-04-26 21:36
    You could look at Chapter 6, activity #6, ROAMING TOWARD THE LIGHT in the Robotics with the Boe-Bot text version 2.2, page 219.··You can download the PDF file right on Parallax's website.· Go to Products>Books>Robotics with the Boe-Bot>downloads and it's right there, can't miss it.· Hope this helps.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    There is always an answer.

    There is always a way.
    There is always a reason.··· -SciTech02.
  • wperkowperko Posts: 66
    edited 2006-04-26 22:21

    cool.gif Hi,

    Reading that section it say's the BoeBot has difficulty working in bright lighted areas ... in fact they suggest changing the circuit and software to look for a hand shadowing one or the other photo-sensors! Hence, this will not work during daylight hours outside and flying an airplane at night doesn't allow for much of a video of the flight!

    QUOTE
    Unlike RoamingWithPhotoresistorDividers.bs2 on page 201, this program will be very responsive to your hand casting a shadow over the photoresistor, regardless of whether the light is bright or dim. This program does not need to change resistors depending on the lighting conditions. Instead, it takes into account the lighting conditions and the sensitivity adjustment is made in software using the average and difference variables.
    end QUOTE

    OTOH, maybe a specific light frequency would work if I can get a bright enough light source on the R/C airplane ... I'm sure if I can solve the tracking of an airplane up in the air, the cars and most other objects will become much easier ...

    But, how do I get a bright enough light source at a specified wavelength the camera system can track? Laser beams are too narrow ... can I make a laser bean that would be extremely visible over bright daylight?




  • Beau SchwabeBeau Schwabe Posts: 6,545
    edited 2006-04-27 05:05
    wperko,

    I have been trying to find an IR camera that we used more than a decade ago when I worked for a prosthetic facility with no luck so far.
    This camera was used to do 3D gait analysis, but this was no ordinary IR camera. It would provide an RS232 output signal containing X,Y
    coordinates for any IR markers (reflective tape or IR LED source) that were detected in it's field of view. The gait analysis software was
    written for a Mac, but the RS232 output from the camera could connect to anything.
    you said...

    ...this will not work during daylight hours outside and flying an airplane at night doesn't allow for much of a video of the flight!...

    Unless pointed at the sun, this camera would work outside as well.


    I will keep looking, and see if I can find something....


    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.

    Post Edited (Beau Schwabe (Parallax)) : 4/27/2006 6:32:47 PM GMT
  • Beau SchwabeBeau Schwabe Posts: 6,545
    edited 2006-04-27 18:32
    Update!!

    Here is a link to the camera that we used... http://www.qualisys.com/proreflex.html



    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • mediaprefectmediaprefect Posts: 31
    edited 2006-05-05 06:49
    Once you have found a "beacon" that can be "seen" by your chosen detector, perhaps you could use a simple baffle between 4 pickups and use the output of the 4 quadrants into differential amplifiers (op-amps) to generate x & y error signals to control you pan/tilt tracking. A more detailed explanation is available here: http://www.spawar.navy.mil/robots/pubs/oceans97.pdf

    Probably way too expensive, but good for ideas, see also: http://www.analogmodules.com/Catalog/quadtrak.htm http://www.irassociates.com/quadrant.htm

    I know I've seen plans for a quadrant detector for solar tracking but I couldn't come up with one for you.
    Then there is the video approach as Beau mentioned: http://www.jrobot.net/ but perhaps more affordable.

    As a tangent, the FIRST competition teams again used video cameras to find a target while being pushed around on the field and shoot balls into the main goal. Goto USFIRST.ORG
  • wperkowperko Posts: 66
    edited 2006-05-05 13:40
    cool.gif Hi,

    I guess I should make the tracking system requirements a little more clear ...

    I want to be able to track something as fast as 50-miles/hour roaming around in an area approximately
    3,000 feet wide, 500 feet high and maybe 200 feet deep.

    I also want to track moving objects in smaller areas too ... maybe down to a square yard ....


    These objects would be moving all the time ... those in the air (R/C Airplanes) might have a little more contrast agains the background at most times but would be subject to changes in color by distance and angle from the cameras.

    I want one system that can track the objects closely enough that one or two other cameras being guided by the tracking system can be zoomed-in for close-up photography.

    The tracking system whould need to be able to move the camera in about 90° wide and 60° tall arcs in about 1-2 seconds ...

    Solar tracking systems using "quadrant" tracking are designed to move in a 90° arc at the rate of 7.5°/hour ... that's about 36,000 times too slow for this project ...

    A "Quadrant" tracking system might work for a very wide-angle lensed camera system ... but would not have enough accuracy to track a zoomed-in camera on a Mack Truck ...

    OTOH, a "quadrant" tracker might be able to guide another quadrant tracker which might be able to guide yet another and another and another ... therefore a fairly well zoomed-in system ... but now the load being moved has increased quite a bit ... and each tracker would have to be calibrated for a specified area of viewing ...




  • Tom WalkerTom Walker Posts: 509
    edited 2006-05-05 14:25
    When you come up with this, the military will be very interested in talking to you...[noparse]:)[/noparse]

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Truly Understand the Fundamentals and the Path will be so much easier...
  • FranklinFranklin Posts: 4,747
    edited 2006-05-29 06:22
    This might be of some help.· http://www.cs.cmu.edu/~cmucam/

    There are several sites that talk of using this for image tracking.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    - Stephen
  • iamscottymiamscottym Posts: 30
    edited 2006-06-02 04:31
    Not to hijack the thread, but I too am looking for some sort of tracking system. I was thinking if I could line up the cmucam2 with a scope that should meet my needs. Would this work?

    Thanks,

    Scott
Sign In or Register to comment.