Real-time motion-sensitive image recognition system

Abstract

This study aims at realizing an effective motion analysis for real time security application by using image processing techniques. After obtaining from the environment, image frames are used to sense and perceive a mobile object, as well as its size and motion degree. In case estimated values exceed its threshold, it is decided whether this image belongs to a human being or a destructive object. Images are stored to the hard disk and then a vocal warning system takes over. The camera is driven by a step motor by which it is possible to track mobile object continuously. The system is enough sensitive to perceive slightest motion and allows user flexibility to select security level. System identifies the person inside by detecting the skin color algorithms and records the image into the computer. The application software has been developed by using C# programming language of. Net technologies.