Video anomaly detection using morphological processes

One of the most important security problems in public places is an abandoned package. It is not practical method that security person is watching security cameras video in order to detect abandoned package. For this aim when public security becomes most important, abandoned package that threatening public security detection system is proposed in this thesis. The proposed method is based on video backgrounds. Objects added on current background are detected by comparing consecutive backgrounds. In proposed method, morphological processes are utilized for reducing noise.