Video Surveillance for Entrance Control

• Author: Lee, Felix, Dipl.-Ing.
• Supervisor: Bischof, Horst, Univ.-Prof. Dipl.-Ing. Dr.techn.
• Supervising institute: Institute of Computer Vision and Graphics

Abstract

We present a method to detect and count people in a video sequence using digital image processing. This method is a part of a control system that is designed to survey entrances of ski lifts. With a predefined installation of video camera system we have to extract the human body, consider its contour and find out the number of people passing the entrance. This is done by the detection system. This detection system uses a hybrid algorithm for detecting moving objects, by combining an adaptive background subtraction technique with temporal differencing algorithm. The number of people can be estimated by finding the significant indentations of blobs. To improve the result we track human inside the entrance area. A tracking system based on the CONDENSATION algorithm is used and described.