Dear BIOMCH-L netters,

I am writing in response to A. P. Bagchee's request for information on
infrared camera lights.

Although he doesn't say so, I assume he is using incandescent IR
lights.

For this use, a better IR source would be LEDs. In our experience,
they consume only about 9 Watts per camera and don't heat it up at all.
They can be left on indefinitely. Since the effective angle of
reflection of the retro reflective targets is very small (less than 1
degree), the LEDs have to be mounted in a tight ring around the camera
lens. Further, if a camera vertical drive is available, the LED
lighting can be strobed to stop the action in each video field and
control the amount of light. Pulsing the LEDs in this way allows them
to be driven at higher currents, resulting in greater efficiency.

Many camera manufacturers use filters on their sensors to block IR
because it tends to smear on the sensor. In the March 1985 issue of
Robotics Age, smear is explained as follows. "When infrared light
enters the CCD image sensor, it passes through the depletion layer and
reaches the neutral region underneath. Electrons are generated in the
neutral region and diffuse into the vertical shift register." This
results in a vertical white bar in the camera view.

Rumors we have heard about different light frequencies follow (but we
have never seen studies proving or disproving any of them.) Does
anyone have hard data on effects or benefits of different scene lighting
frequencies?

Infrared safety: it is said that people can suffer retina burns from
looking into IR sources they didn't know were on.

Visible light: it is said that some animals and children are distracted
by visible camera lights. People who use visible red LEDs report that
this is not a problem in practice.

I hope this helps you in your lab setup.

Ken Landaiche
H/W Mgr
Motion Analysis Corp.
(ken@mac.com)