I have a project that I am playing with for an e-learning development tool for our school's faculty. We had started playing with Augmented Reality; and while I think there are still too many limiting constraints on that technology for us at the moment, it did give me an idea of integrating papervision 3D with a webcam for lectures using a more conventional approach.
What I would like to do is to place two layers in a canvas and record the canvas as a camera source to send to our FMS server. Layer 1 would be the Video object attached to the webcam, while Layer 2 would be a papervision viewport in which the instructor could drag/drop and interact with 3D models. This seems like it should be a fairly straightforward implementation (theoretically). Using the Bitmap class I can capture a snapshot of the canvas (both layers). Capturing 30 bitmaps a second would be similar to the camera's operation (without the compression). The only issue is that NetStream has only an attachCamera and an attachAudio method. I cannot find any Open Source code for flash.media.Camera or flash.net.Netstream so I cannot see how netstream is pulling the info from the camera or how the camera stores the captured video information.
Is there a way to do this using Flex or AIR????
it might be possible with AIR FP10 and Alchemy that someway taking input from any visible abject on stage and wrap it into camera Stream and then pass it the NetStream , but i am not sure . you may ask this question in alchamey section too.
Was hoping that it would be simpler than trying to solve it in C/C++; perhaps extending the Camera object or some obscure documentation on protected components. Even NetStream has attachCamera(x:Camera, t:Number) where you can throttle the number of snapshots that are transmitted to the server with t.