In this paper the challenging problem of unusual event detection in video surveillance system is considered. In real time applications, tracking target in low resolution(LR) video is really a challenging task. The existing methods require high computational cost because it uses super resolution techniques for the enhancement of LR (low resolution) video. Proposed algorithm require low cost and is fast enough because it process low resolution frames and it is able to recognize the occurrence of uncommon events such as overcrowding and fight in the low resolution video without using any classifier and training datasets initially. The application of this proposed method is to enhance the ATM security.