Automated Video Surveillance at Night

Visual surveillance systems have been used for decades to deter and detect crime by providing extra sets of eyes to security personnel. However, using human surveillance operators to monitor banks of video displays is problematic for many reasons, including operator fatigue, complacency and the inherent difficulty of detecting all objects of interest when many video streams have to be monitored simultaneously. This monitoring task is made even more difficult at night since nighttime video data can be relatively noisy, or lack the amount of information, color and texture for example, that is often available in daytime video. In order to assist security personnel monitoring a site at night, we propose to extend our Phase I intelligent video analysis system for alerting security personnel when something of importance appears within the video. Working with the operator, this component intelligently removes motion clutter, detects and tracks objects, and using a definition of interesting behavior from the operator, determines if an alert is warranted. In order to deal with the characteristics of nighttime video, the system exploits motion within the video directly and uses how an object moves, rather than its structure or static appearance, in order to analyze objects and their motion.