Automated Intelligent System for Recognising Objects on Scene

Aim of the book: To build an efficient automated recognition system for recognizing objects on a scene, so many problems must be solved and many assumptions must be made. These assumptions e.g. include the image has two grey levels (using a threshold filter for 256 grey levels), the object surfaces are smooth and there is no shadow. The problems to be solved are; size changing, translation, rotation, reflection and noise effects as well as the deformation of shape. Automation system was implemented by applying two approaches techniques; one is the application of the invariant moments. The other one is the invariant auto correlation. Although the invariant moment approach was used, it was sensitive to noise and useable with the video images. But the invariant auto correlation approach solved the sensitivity to noise.