Imagine a workshop ceiling which is dotted with small optical PTZ mounts on a grid - say, every 30 to 50 cm. But instead of a camera imager at the focus of the zoom lens, we have a bright LED. To get decent performance, we'll want the lens to be much larger than the ones used for cameras also - my guess is that somewhere between 30 and 50mm is about right. The "zoom" design need not be sophisticated. The idea is for each mount to be able to create a spot of light between 15cm and 1m in diameter right in front of the user, or the user's face, or where the user's hands are spending most of their time...
In addition to cheap, simple, moving-head lights (the idea was developed on a larger scale for stage lighting in the 1980s) we will need some kind of sensor system to detect not only the position of individuals (plural!) in the room, but which way each one is facing. This appears to be do-able with simple cameras feeding a low-end processor running a face detection algorithm. We probably need a minimum of one camera on each wall with multiple cameras for especially long walls.
Ideas here? Anyone? Is there a better way to do this? There will be intellectual property disputes over this, and in the end only one or two systems will be truly elegant, the others will be lawyer-evading kludges.
Ahem: nothing which requires the user to wear a tracking device is elegant. Ahem. I have spoken.
Once we know how many users are in the space, and which way each one is facing, we can tell each lighting unit which way to point, always coming from above, the sides, and behind a user to illuminate whatever they are facing or working on or looking at - never shining toward their face.
Some units can be dimmed or turned off entirely or turned into indirect/accent lighting by facing away toward a wall when they can't be put to use for a user.
With a sufficient number of units (more than 256, I will argue) the illumination becomes effectively shadow-free. The small office I am sitting in is eleven feet by twelve feet. Using a 1 foot grid spacing in this room would require ninety lighting units.
The appropriate size of each spot might be determined by a remote on prototype units, by voice command on early shipping units, and by AI (using "busy area" image analysis to determine the size of the work space) on mature systems a few years in.
Someone is almost certainly working on this right now. There are probably patents on some of the math required. Someone may even have applied for an overall "basic" patent on the general idea of an intelligent, task-based, person-aware, lighting system as I have just described.
No comments:
Post a Comment