29-12-2012, 03:04 PM
Stereoscopic Vision based Navigation
1Stereoscopic Vision.doc (Size: 29 KB / Downloads: 23)
Abstract:
Older obstacle avoidance systems use very delicate and heavily prices sensors like RADAR and LIDAR (Light Detection and Ranging), these sensors are very accurate but lack the robustness due to numerable moving parts. The aim of the project is to use low cost and readily available hardware to design a system which calculates the distance to any particular object. The final sensor would be self contained, would have lower number of moving parts and would use software techniques for calculating distance making it more robust.
The system would consist of 2 web cameras spaced a specified distance apart, connected to a PC running the distance calculating algorithm. This system could be used in collision avoidance systems, auto navigation systems etc as a much more cost effective replacement for RADAR and LIDAR.
The stereoscopic vision based navigation is a system which contains 2 cameras at a specified distance which will calculate the distance from a specified object. This algorithm would be used to trace paths and navigate obstacles.