Automatic event detection in CCTV surveillance

Using a simple but robust real-time algorithm that allows tracking of multiple objects in complex environments, the authors propose an application in automatic surveillance by linking blobs’ information, in terms of trajectories and positions, with the events to be detected. Working in the background of the CCTV surveillance systems and based on whether a selected event has been detected or not, such an application can decide which areas should be monitored, helping the operator by avoiding empty scenes and showing him footage more likely to contain abnormal situations. The paper shows an example of characterisation of events to be detected in transport environments.