CMUcam II and competition lighting environments

[dlavery]

As you experiment with the CMUcamII camera from the Kit of Parts, make sure that you pay attention to the lighting environment in which you test. Most, if not all, of the competition venues will be very well lit, and the light levels will be much higher than you might expect. This will have an affect on the illumination levels received by the camera. But perhaps even more important, the type of lighting used in the competition venues will significantly affect the response of the camera. If you are testing your camera in a lighting environment that is markedly different, you may want to take steps to adjust the lighting to be more closely aligned with the competition lighting.

Back in my old stage lighting days, I learned that most gymnasium and coliseum locations use sodium vapor or mercury vapor lights. These lights usually have a color temperature of around 2000-2100 Kelvin, and have a noticeable yellow tone. This yellow tone will shift the color of the vision tetras as they are perceived by the camera. As most of the competitions will be held in these types of lighting environments, you will need to be aware of, and compensate for, this warmer lighting.

If your development environment is lit with standard florescent lighting as is found in most school rooms and shop environments, be aware of the heavy blue bias in the lighting. Florescent tubes have a color temperature of around 6500 Kelvin (standard broken spectrum tubes). This is noticeably bluer/cooler than typical daylight at mid day (around 5500 Kelvin) or bright metal halide lighting (around 4500 Kelvin). If you tightly calibrate your camera for this type of environment, you may run into problems when you get into competition lighting.

You can spend a lot of money to get a lighting colorimeter or spectroradiometer to measure the precise color temperature of your environment. But there are less involved, and less expensive, ways to get a reasonable approximation of what you can expect in the competition venues. One of the best ways to compensate for the differences in the lighting environments will be to get a 500 watt halogen work light (available at Home Depot and similar locations) and use it to light your test environment. Halogen lamps have a color temperature around 3200 Kelvin. One or two of these will dump a lot more lumens and a lot more warmth into your lighting. Add one or two targeted incandescent tungsten lamps (color temperature 2850 Kelvin for a typical 100 watt bulb), and you will be getting close to where you want to be.

-dave

[Billfred]

Just as an idea--is there a way to set up at the venue a "Tetra-Rama" near the field where teams can calibrate their CMUcam with the honest-to-goodness tetras in the lighting of the arena?

Sometimes there's just no substitute for the real thing.

[russell]

I am guessing that they are going to do just that. I also dont really understand how the cameras work but couldnt they just post the values on the field? Or does it vary between cameras?

[Adam Krajewski]

Originally Posted by Billfred

Just as an idea--is there a way to set up at the venue a "Tetra-Rama" near the field where teams can calibrate their CMUcam with the honest-to-goodness tetras in the lighting of the arena?

Sometimes there's just no substitute for the real thing.

Calibration numbers will be provided by FIRST per this post: http://www.chiefdelphi.com/forums/sh...0&postcount=19

[dlavery]

OK, so I asked Steve Sargeant (the guy responsible for the lighting at all the regional competition events) exactly what the lighting set up will be at the events. He just responded that they are planning to use quartz halogen light arrays at the events, at least 85 foot candles, with a color temperature of 3200 Kelvin. So the original recommendation to use a pair of 500 watt halogen lights to set up your development environment to test the CMUcam II is still good. It should put you right where you want to be in terms of the typical lighting that will be at the regional competition.

-dave

[lynca]

The lighting is definitely difficult to test. We test with mercury lights which produces a huge amount of IR. The basic camera tracking would not work out of the box. We had to play with white balance on the configurations of the camera to tune for our lighting conditions. Be careful, the CMUcam GUI states white balance is turned on, but updating the configurations is not always assured depending on the camera state.

116's modular electronics is awesome!
if you look closely at our controls box released at the start of the month, you will see a similar modular electronics box with pneumatics included!!
soon we will upload pictures,
~Andrew Lynch