SuperVision: Video Content Analysis Engine for Videosurveillance Applications [chapter]

Lisa Usai, Francesco Pantisano, Leonardo G., Franco Selvaggi
2011 Video Surveillance  
Introduction Greater challenges in the security area and declining cost of technology have promoted the development of ever more sophisticated video surveillance systems. Such systems are widely employed both in the public sector, to support police activities for example, and in the private sector, in banks, shopping centres security, etc, working 24 hours a day, 7 days a week. Beside security applications, video surveillance is successfully employed in other fields, such as monitoring traffic
more » ... r studying people's behaviour or consumer's preferences. The increasing extent of the areas to be monitored requires the use of a large number of cameras. Video-streams flow to central control room and are displayed in real time to operators. The large amount of data makes the task of security staff demanding but also very tedious. Although security operators are trained, it's impossible they maintain high levels of attention when confronted with multiple inputs for more than a few minutes, (also because most of the time video streams show ordinary behaviour). Furthermore sociological researches (McCahill & Norris, 2003; Smith, 2004) have proven that often it is the operator who decides on which camera to focus his attention, basing the decision on the appearance rather than the behaviour of people in the scene. Video content analysis represents a solution to these problems. Its main purpose is to analyze video streams and alert the operator only when relevant events are detected. This will help solve the problem of operators discontinuous attention. Even the European community is paying close attention to these issues and in recent years several funded projects were launched to develop the most appropriate technologies to solve specific problems. The main goal of ISCAPS, for example, has been to reinforce security for European citizens and to try to reduce terrorist threats. The aim of SAMURAI is to develop and integrate an innovative intelligent surveillance system to monitor people and vehicle activities in critical public infrastructures and their surrounding areas. SUBITO addresses the problem of automated real time detection of abandoned luggage, fast identification of the owner and his/her subsequent path and current location. The organization of intelligent video surveillance systems is hierarchical and generally starts with object detection, estimates the position of the detected object over time (object tracking) and describes what happens in the scene (event recognition). Video Surveillance
doi:10.5772/15679 fatcat:qde3mn7cqjfojp3enlaegain4u