They create a model to analyze images from video surveillance systems in public places

Engineers from the Higher Polytechnic School of the Autonomous University of Madrid (UAM) propose a new model to develop automatic video surveillance sequence analysis systems. Unlike the current, These systems could be adapted to the particular characteristics of a certain scenario or situation..

Engineers from the Higher Polytechnic School of the Autonomous University of Madrid have created a system to analyze the images captured by video surveillance cameras and that, unlike classic models, adapts to the particular characteristics of each scenario or situation.

Video surveillance systems allow a specific area to be monitored through the use of several cameras connected to a monitoring center.. They are currently in high demand systems to detect potentially dangerous situations in public places such as airports., subway stations or sports venues, and that require specialized people to review the images.

To make supervision more effective, Automatic video analysis tools are currently used to focus attention on a specific monitor. (with images of a potentially dangerous event, For example) or perform efficient searches on recordings, among other uses. This need has turned the automatic analysis of video surveillance sequences into a very active area of ​​research..

Current automatic video surveillance systems are based on three-stage analysis: detecting objects or people of interest, tracking and extracting features from these objects or people, and detection of unusual events (abandonment of objects, entry into prohibited areas or acts of vandalism).

These stages work independently of each other, and apply in the same way in all scenarios and situations. The engineers at the Higher Polytechnic School have warned that there is a dependency between these three stages and that their application can and should be adapted to each type of situation., given that, For example, A potential robbery in a lonely parking lot does not require the same detection techniques as another in a subway station with a high density of moving people..

The video surveillance model they propose therefore has the ability to adapt the analysis stages to each scenario and situation., and focus attention on cameras with greater data complexity.

View test results