Creating basic SLAM system using two webcams and OpenCV

edit

Hello everyone,

I have a very simple project; I am attaching two webcams to my robot, and I need to calculate optical flow for XYZ.

To prototype the program, I built everything in Unity. To begin, I've taken a 3D scan of the interior using Scenect. I've converted it into a 3d model, and loaded it into unity.

Next, in part of the 3D scan, I've placed a Barcode using OpenCV to detect the initial position in 3D space, (this starts the program as well), using Marker tracking.

Once the program is initialized, I'm gathering optical flow points from two webcams; the first triangulates XY movement using the front facing camera, and the second calculates flow at a 90' angle; or XZ optical flow.

In the scene, I've positioned webcam textures that are using the optical flow unity example by Equinox software. It creates red "bullseyes" on tracking targets. These textures are world-space canvases, and are parented to the main camera, (the eyes of my robot).

Finally, I have target points parented to find red target symbols on the webcamtextures.

For standard movement, I'm using a high precision gyroscope & accelerometer. To correct for sensor drift, I am firing traingulating beams at the targets in the webcamtextures. The rays will collide with the collision mesh, which has been simplified to help with tracking. The unity beams will find the closest adjacent points, and correct camera placement in the 3D space.

I'm nearly done with everything, except writing some of the code. Can anyone help me with this? I can compensate for work, and I really need to get this done as soon as possible. I'm super busy with other aspects of the business, and can really use the assistance. Serious inquiries only please!

My team uses skype to chat.