Need to Build a Rotating Camera Device

There are all sorts of options here. Presumably, the unit is to be mobile? Perhaps run from a laptop or similar?

Either servo or stepper options are viable. One option would be to look at Labview & either USB based control or perhaps compactRIO or single board RIO depending on the complexity/cost you're looking for. Go to ni.com here and root about.

Of course you could just buy one... like this for instance

Or, if money is no object, you could invest in;

Catadioptric cameras

Lens and mirror based (catadioptric) cameras consist of lenses and curved mirrors that reflect a 360 degree field of view into the lens' optics. The mirror shape and lens used are specifically chosen and arranged so that the camera maintains a single viewpoint. The single viewpoint means the complete panorama is effectively imaged or viewed from a single point in space. One can simply warp the acquired image into a cylindrical or spherical panorama. Even perspective views of smaller fields of view can be accurately computed.

The biggest advantage of catadioptric systems is that because one uses mirrors to bend the light rays instead of lenses (like fish eye), the image has almost no chromatic aberrations or distortions. Because the complete panorama is imaged at once, dynamic scenes can be captured without problems. Panoramic video can be captured and has found applications in robotics and journalism.

GCPad, I could of course buy one but it needs to go on some sort of a stand, and it will need to be modified a bit to for example take a vertically -90 shot. And these things seem to be too expensive, too heavy, and much more complicated than what I need. What I would like to do is try to make a simple machine myself, unless just buying something and putting it atop a stand would be a better idea? thanks for your tips. Much appreciated.

34point5, that is a very interesting idea, I didn't know about them. Must look up more info about these cameras. thanks

Here is a product I have used in building robotic camera prototypes.

http://servocity.com for pan tilt heads and servos. Reasonable prices. Wide variety.

How are you planning to trigger the camera?

http://pro.sony.com makes FCB block cameras that can be triggered by the code from the atmel processors (VISCA protocol)

Yes, steppers would probably be better. Precise positioning without the need for validating the position for the precision you'll be operating at.

Google "Arduino stepper motor" to get tons of ideas about using open source arduino and its free programming environment as a controller for you system. There's tons of freely available code on the net and at the Arduino home page.

I'm currently building a 3D (3 axis) printer with steppers that's based on Arduino control.

Hooker

Ummmm - Hooker, no offense, but the perpetual problem in film motion control, with steppers, is they might not step, due to the mass and balance variables of camera bodies and lens choice - and you don't know it hasn't stepped until "post".

Also, if he wants to do a pan in discrete increments, he has to match the stepper increments (by whatever ratio) to the chosen lens 'acceptance angle', or overlap and "fix it in post".

It can be argued (as many motion control failures have) that 1.8⁰ is adequate resolution geared sufficiently for long lenses - but steppers loose resolution and torque with rpm.

Thus, as things become possible, film wants more, like faster pan (higher rpm) and longer lenses (finer increments). At this point steppers run out of combinations and the entire control system is scrap.

So ok, steppers look good on paper, but nearly all modern remote heads use encoded brushless motors because of "variables" in - well - just about everything you can poke a stick at.

Taken another way- steppers are "reinventing the wheel" like the original UK "Hot Head", which is circa 1980, and so 'old hat', I can't even find a web reference.

Similarly, steppers are long gone from CNC, bar foam cutting toys.

Then again the OP is doing it in wood maybe you're right.

Good points, all, 34.5, and no offense taken.

But I stand by my recommendation. I was thinking mostly in terms of the mechanics and a lack of seeming concern about cycle or process time, and not taking into consideration recording "movies", just discrete snapshots. And the implication that money is no object but technical ability to throw together a complete system may be lacking.

Anyway, as one who used to breadboard systems like these for a living (research machinist at NASA) I don't see a big deal to put together a hefty enough mechanism to handle torque and position requirements for moving a small to moderate sized camera around with relative precision. The arduino system is robust enough (programatically) to aim the camera, with whatever lens, per the parameters described. Multiple Arduino programs could easily be made available and uploaded to accommodate the desired lens focal length/"acceptance angle" for a discrete session.

I will absolutely grant you that encoded motors with feedback validation would be the cat's meow but I didn't feel that the OP's original description of the task warranted that. Maybe they'd be Revision B once the initial mechanism is proved? But that's what design review sessions are all about, right?

Hooker

Yep - and its all a bit 'open' when you don't know all the other factors like payload and operating parameters - or even where the OP is (so the operating environment, through to which suppliers one might point to).

It might be equally valid, or more so, just to tell him to buy a small rotary table or indexing head.

Friends,

A small factor that should be added to the mix: Let us assume that the first order approach has 13 horizontal increments, each with 11 vertical increments (totaling 143 images). When the vertical angle puts the camera lens parallel to the horizontal axis, each one of its 13 pictures centers on the same space, just at 13 different angles of rotation. In writing the code or program to move the camera, a second order approach would change the number of horizontal increments, with it approaching 1 (or 2) as the vertical angle goes from 0 to either +90 or -90 degrees. The entire area could be covered with about 88 images (and maybe less).

--JMM

Maybe something like this could be adapted reasonably

http://photoshipone.com/ps1-products/