Alright thanks guys for all the comments.
So from what I gather, how well the sensor will work depends greatly on the type of glass I'll be using. As someone else is manufacturing the housing for this proximity sensor at a later stage I can't really be sure of how it will work I guess.
I think I'll just have to experiment with as many different types as I can.
I'll give a brief overview on the project in case anyone has any other thoughts or suggestions.
There exist a number of rectangular panels, formed with steel and a glass surface on one side: 1.2m x 1.2m x 20cm. They will be used to build a wall. Inside each of the panels I will be placing a grid of high power LEDs. The people in charge of the project would like subsections of the panels (mini LED grids) to light up as someone walks past that section of the wall.
The problem is that I'm not to place anything on the exterior surface of the wall, or to cut out any sectons of glass, etc. So the sensing has to be done within the panel, through the glass.
And of course, the budget for such a project is minimal. At most I'll get about US$700 for 500 LEDs and all the electronics, including power supply. Right now I'm working on making 36 sensing areas within the wall.
I don't need any accurate distance measurement, just a "presence" of someone close to the wall. 1m outside the wall would be ideal, but if people need to brush past, basically touching, the wall then that will have to do.
I've been playing around with my own configurations of IR emitter and detector pairs, but the various levels of daylight aren't making for a stable output.
I know the SHARP sensors have some subsidiary electronics which help negate the effects of ambient light, which is why I'm thinking of using them.
There does exist a cheap SHARP presence sensor which claims to work up to 400mm - GP2Y0D340K @ $6.50 which just might work for me.
I guess I'll just have to play around with a few to see what I can get with different types of glass.