I. Introduction
In recent years, everyone is conscious about security issues and it has become one of the basic needs and necessities of life. Due to increase in number of terrorist attacks, there is a need to develop a frame work for intruder detection in order to generate alarm whenever an intruder crosses a boundary line and enters into a restricted area. The primary concern of this paper is to propose a human intruder detection mechanism. in order to manage and monitor the security problems, i.e., tracking of multiple objects in an automatic way without the interference and involvement of human effort. In the proposed method, human intruder motion is detected by background subtraction [1]–[5] and tracking is done with the help of Kalman filter [2]. The assignment problem between Kalman predicted tracked position and its assignment to the individual object is tackled through Hungarian algorithm [4]. The algorithm will be able to distinguish and classify between human and nonhuman efficiently. For better results, noise is removed with the help of morphological operators. The proposed method works with daylight as well as night vision or thermal imagery. Rest of the paper is organized as follows: Section II explains the proposed method, Section III discusses results, and Section IV highlights the summarizes the proposed work and suggest future directions for it.