Stereoscopic Video Surveillance System
No Thumbnail Available
Date
2011-08-12
Journal Title
Journal ISSN
Volume Title
Publisher
UMT Lahore
Abstract
The aim of the project is to design and implement a stereo vision system using two simple web cameras to simulate the stereopsis of human vision, which allows for accurate depth and distance perception from a pair of two-dimensional images, and a pair of small servo motors for pitch and yaw.
For the purpose of this project, we have two standard webcams and two servo motors mounted on a mechanical platform. Motor control is performed via a PIC microcontroller, which itself is controlled via a serial interface with MATLAB. Image processing, tracking, and the implementation of stereo vision will also be done in MATLAB, where existing algorithms will be optimized and new ones utilized. The focus here will be on reducing computation time for the stereo image pairs, since image processing is a CPU-intensive task and can take a while on the processors found in today’s laptops and PCs.
The true scope of this project and its many uses is better understood by observing the burgeoning field of computer and robot vision, which boasts technologies such as early obstacle warning systems in vehicles and intelligent security systems, both of which utilize stereo vision to make detailed depth maps. The now-defunct Mars Rover is a pioneering work in this field and is still heralded as one of NASA’s greatest achievements to date. This BS project attempts to mimic as much of these cutting-edge technologies as possible.