In this paper, an effective algorithm which can detect moving objects from the image sequences and classify an intruder from other moving objects is presented for an implementation of unmanned surveillance system.
The proposed algorithm uses a stic...
In this paper, an effective algorithm which can detect moving objects from the image sequences and classify an intruder from other moving objects is presented for an implementation of unmanned surveillance system.
The proposed algorithm uses a stick figure model for features of the moving object since the stick figure model can be represent the characteristics of the moving objects.
The stick figure model is extracted by applying the thining algorithm to the binary difference image between the reference and the image of a moving object.
The computer simulation is carried out to analyze the performance of the proposed algorithm. The results show that the algorithm can discriminate a moving person and a moving animal with high degree of reliability.