Abstract

The task of constant monitoring of video streams from a large number of cameras and reviewing the recordings in order to find a specified event requires a considerable amount of time and effort from the system operators and it is prone to errors. A solution to this problem is an automatic system for constant analysis of camera images being able to raise an alarm if a predefined event is detected. The chapter presents various aspects of video signals processing dedicated to detection and monitoring of threats in urban areas. First the video analysis meth-ods are presented, then recognition algorithms are introduced for detection of im-portant events, and finally obtained results are discussed. The section 2. is dedicated to basic video analysis algorithms aimed at detec-tion, tracking and classification of moving object appearing in a camera field of view. First in section 2.1 moving object detection is presented. It employs back-ground modelling and subtraction for determination of non-stationary objects de-tection. Then the problem of objects shadows is discussed, and finally a stage of image segmentation is presented. In Sec. 2.2 moving object tracking is described. This process entails a necessity of solving many problems regarding object occlu-sions. An approach utilizing changes in time of object state described by Kalman filter is discussed, which concludes low level processing of visual media. An outcome of low-level processing serves as a input to higher-level analysis, where it is determined whether they are actually dangerous situations. The high-level video processing is presented in Sec. 3. It starts with object classification (Sec 3.1), and then event recognition is performed (Sec. 3.2). Modern video sur-veillance systems contain multiple cameras, covering a wide area, some of which are able to pan, tilt and zoom view area (PTZ cameras). A technique for position-ing of PTZ cameras is presented in Sec. 3.3, allowing tracking of moving object.

Citations

Authors (4)