Computer Vision refers to the use of cameras and other input devices for computers to see like human. Human needs the brain to interpret whatever images are captured by the eyes. Similarly, computer needs software to interpret camera images.
A modern video surveillance system deploys IP cameras to capture real life activities and a Video Management System (VMS) to interpret the images for human. This system is already doing computer visioning although the context is mainly in physical security surveillance.
Computer Vision is applicable to many walks of life than just security surveillance. By recognising pattern of images, computers can tell if the pattern is a person, vehicle, or an object etc. The algorithm is simple but computation can be tedious and accuracy would be doubtful.
Data modelling will be an ultimate means of analysing sequential static images.
The University of Auckland multimedia imaging website has this short introduction (www.mi.auckland.ac.nz ): Computer vision aims at understanding or modelling static or dynamic 3D objects or scenes based on captured image data. Typically it requires an understanding of projective geometry (how 3D objects are mapped into one or several 2D images, possibly from different perspectives), and combines (static) picture analysis with image sequence analysis. Image sequences are captured by normal video cameras or more advanced specialized cameras.
Return to Technology Park