Camera rotational stabilization

Hakhak

Hi there,

I'm doing a project on camera stabilization, and I'd like to know how many degrees of tilt, yaw, roll can be stabilized. Olympus, for instance, has 5 axis stabilization systems (sensor based) of which three rotational (pitch, yaw, roll). Does anyone know how much this sensor can rotate? And how much is theoretically possible - for in-lens and in-body, and perhaps a combination of both?
Or perhaps someone could give me nudge in the right direction of ways to calculate this?

Many thanks!

ziggy53

I suggest designing an empirical test on an apparatus which can move in single axis of movement first, and then add more types of movement to test simultaneous motion stabilization. Basically you want to test both motion rate and motion limits.

A stacked piezoelectric actuator may be needed to provide the necessary action for each type of motion. I suggest building the mechanical rig first, to both hold the camera body and to provide the required motion control, and then match the piezoelectric actuator specifications to the necessary load and degree of motion for the rig plus camera body.

I suggest using a pinpoint of light as the visual target for the camera/lens, and that should allow an image with a quantitative visible deflection, indicative of the success of the stabilization. Once you know the limits of stabilization, calculating angular deflection should be relatively straightforward.

As for the theoretical limits for chip based stabilization, I suggest that different applications might have different requirements. A true theoretical limit for any system would have to include a definition and description for size and mass for all components. Computer software related to FEM and FEA should be a tremendous help.

Brett1000

ziggy53 wrote: »

I suggest designing an empirical test on an apparatus which can move in single axis of movement first, and then add more types of movement to test simultaneous motion stabilization. Basically you want to test both motion rate and motion limits.

A stacked piezoelectric actuator may be needed to provide the necessary action for each type of motion. I suggest building the mechanical rig first, to both hold the camera body and to provide the require motion control, and then match the piezoelectric actuator specifications to the necessary load and degree of motion for the rig plus camera body.

I suggest using a pinpoint of light as the visual target for the camera/lens, and that should allow an image with a quantitative visible deflection, indicative of the success of the stabilization. Once you know the limits of stabilization, calculating angular deflection should be relatively straightforward.

As for the theoretical limits for chip based stabilization, I suggest that different applications might have different requirements. A true theoretical limit for any system would have to include a definition and description for size and mass for all components. Computer software related to FEM and FEA should be a tremendous help.

quite the experimental test !

Hakhak

ziggy53 wrote: »

I suggest designing an empirical test on an apparatus which can move in single axis of movement first, and then add more types of movement to test simultaneous motion stabilization. Basically you want to test both motion rate and motion limits.

A stacked piezoelectric actuator may be needed to provide the necessary action for each type of motion. I suggest building the mechanical rig first, to both hold the camera body and to provide the required motion control, and then match the piezoelectric actuator specifications to the necessary load and degree of motion for the rig plus camera body.

I suggest using a pinpoint of light as the visual target for the camera/lens, and that should allow an image with a quantitative visible deflection, indicative of the success of the stabilization. Once you know the limits of stabilization, calculating angular deflection should be relatively straightforward.

As for the theoretical limits for chip based stabilization, I suggest that different applications might have different requirements. A true theoretical limit for any system would have to include a definition and description for size and mass for all components. Computer software related to FEM and FEA should be a tremendous help.

Wow! What an elaborate reply! Thank you!
I was wondering, if it is possible as well, to get the information through a calculation.. The question is how far can the light sensor rotate in relation to the lens without losing sight? if it rotates 90 degrees it won't catch any light, but between 0-90 what is the max? To continue my project I have to prove that 6-7 degrees would still be possible.