tags:

views:

100

answers:

3

Hello,

my problem is to cover 100% of the room with cameras in 2D. The camera can be modelled as a polygon. I am finished with all the modelling and necessary tools. My problem is with the algorithm. The complexity of blind search for a big room with many obstacles will be too high.

So does anyone know a better (or optimal) algorithm?

A: 

If you do not need to consider obstacles, use a single 180 degree fish-eye lens mounted on one of the walls.

RedFilter
Won't work if there are obstacles.
Mark Byers
i can use even 360 degree if you see one of those, but the user enter the specification of the cameras to be installed
Bass
+1  A: 

This sounds promising:

http://www.springerlink.com/content/u7gp44hchuhly6yg/

Jens Schauder
A: 

Use two 180-degree cameras on walls that are 90 degrees to each other, assuming a room that is a square or rectangle. With that, you get a redundancy for most of the room except for possibly the one corner that falls in between the cameras.

SDGator
the specification of the camera and room are dynamics and entered from the user, so i can't assume anything.
Bass