Systematic Evaluation and Comparative Analysis of Moving Object

Abstract

Object detection is the process of finding instances of real-world objects such as faces, bicycles, and buildings in images or videos. In an automated video analysis it is the fundamental step.An object detector requires manually labelled examples to train a binary classifier, while background subtraction needs a training sequence that contains no objects to build a background model. In order to achieve better performance object detector without training sequence i.e motion based methods came into existence. But when comparing situations such as non rigid motion and dynamic background, motion based object detection becomes a failure. So here an efficient algorithm is discussed which will integrates object detection and background learning into a single process of optimization. It can work effectively on a wide range of complex scenarios. An efficient alert system is implemented so that the moving object can be detected in real time.