Please make sure you recalibrate your CMU sensors whenever you switch from differnt lighting schemes. Anyone who has taken photography knows that diffent lights will actually cause a color shift. Not correcting this colorshift will result in an inaccurate representation of colors which is interesting to talk about in another topic because there is contraversy as to what color Mars is. Flourescent light usually has a green tint, halogen results in a redish tint, and I don’t know what sunlight does.
ah. that’ll be nice to know. :]
thanks!
~Stephanie
(and thanx the rest of our programmers down here too)
Thanks for the “head’s up”!
I had a hunch this new sensor sounded to good. I have not seen it yet, is it diffucult to calibrate or is it possible to progam it for a range of many shades?
You would be suprised how much the CMUCam already has integrated into it. Calibration routines are automated, and there is a fair amount of documentation on the site on how to deal with different lighting conditions. The camera routines that would are dependent upon the lighting conditions are integrated into the camera for the most part anyways.
In short, I think that the lighting will be less of a problem than everyone is making it out to be. Because so much of the camera interface is high level results of very complex, optimized, and tested machine vision routines, I think you will find the output that actually arrives to your robot is pretty clean as long as the lighting is fairly even and bright enough to illuminate the field.
I haven’t tested it though with the IFI routines, so I don’t know. I have, though, seen a CMUCam in action and was very impressed.
When you initially set the camera, you must calibrate it to find what the RBG value sets of the objects you are trying to find. Then, you take the three values, Yellow, Green, and Red, and copy them into the “camera_init()” function in the user_routines_ddt.c file. This is, of course, if you are using the code supplied by IFI. I highly suggest you use this code, making modifications to it as needed.
When you turn the camera on, it will auto detect lighting conditions, and adjust the AEC (automatic exposure control) to compensate for this. In easier terms, it sees how much the light is “grayed” compared to set code values. If it is darker, it allows more light to enter the camera before taking a frame, and vice versa.
IF this turns out to be a big issue, you will need to be careful calibrating at the event. Usually, the pits are lit by venue lighting, normally HID, which has much more blue in it. However, the field is normally lit with theatrical lighting, which is a tungsten source, and has much more red in it. I do not think this will be much of an issue though, as you are looking for the three primary colors (plus yellow). The camera should be able to discern between these colors, even if the calibration is off by a little.
Ehhh… I was only mentioning this because of personal experience and from what I read about the original CMU cam and as experience as a mediocrae photographer. I have about fifty pictures that are completely washed out in a sickly green because I didn’t know that flourescent lighting does this. 
Funny how we computer folk often forget that yellow, not green, is the third primary color. 
Actually, when referring to light, red green and blue are the primary colors, in Pigment, the three primary colors are Cyan, Magenta and Yellow.