I am working on an idea to have a closed-loop automatic projector calibration system. With either a digital camera or camcorder feed images of the projected screen into a computer that with custom software determines for instance if the brightness should go up or down.

I understand that there is a lot of hurdles to overcome here but I was just wondering what is the best way to take the picture of a screen in a very dark room. The pictures I have taken so far with my digital camera haven't turned out very good and definitely doesn't reflect the colors that I see when I watch my projector. Does anyone if it is even possible to take some type of picture (or have a camcorder filming the screen) that would be close to what the eye is actually seeing.

Appreciate all the feedback I can get

Thanks,

You certainly will have an uphill battle trying to calibrate in this manner as the CCD imager in the camera which you are using is not calibrated to a reference level of Gray. There is also the issue of flat field uniformity of the detector and a host of other issues most of which you will not have any control over. The calibration is possible to do in this manner with a high quality CCD camera made for machine vision applications which would still need to be calibrated to a reference level to provide accurate data for the measurements. You will most likely not be able to get very good low light level response either (with a conventional camera) as the detector signal level will be buried in the ambient noise level of the detector as it most likely will not have the dynamic range and required signal to noise ratio needed for this type of application.

Thanks for the feedback. I thought that my biggest hurdle would be to have my software figure out where in the picture taken by the camera/camcorder is the test patterns located. That is non-trivial because of perspective distortion, various similarity transforms etc. I was able to come up with a mapping algorithm that is pretty accurate but I guess if the picture used as input isn't representative of what is truly displayed then this system will not be much good.

Thanks for the feedback. I thought that my biggest hurdle would be to have my software figure out where in the picture taken by the camera/camcorder is the test patterns located. That is non-trivial because of perspective distortion, various similarity transforms etc. I was able to come up with a mapping algorithm that is pretty accurate but I guess if the picture used as input isn't representative of what is truly displayed then this system will not be much good.


What you are discussing is what William Phelps has been doing with JVC projectors to calibrate gray scale uniformity. We are both astrophotographers so I know something about the equipment he is using. You should look at a chilled SBIG astro camera if you are really serious about this. Many of these units are designed for B&W and would need to have color filters automatically rotated into the field for each color. A separate device would be needed to calibrate the gray scale for these.

Thanks.
The whole idea behind is was to create a system where you would use an off-the-shelf camera or camcorder together with my custom software to easily calibrate your projector. As an alternative to buying Avia, DVE HD etc... and mess with your settings yourself. The idea was that the computer would be talking to the projector via a serial cable (for the projectors that have a RS-232 command interface) and based on the picture input my software would tell the projector what settings needs to be tweaked. The idea was more of a consumer product but it sounds like the equipment needed for this invalidates my idea.

... The idea was more of a consumer product but it sounds like the equipment needed for this invalidates my idea.

I would agree.