Real-time human activity monitoring exploring multiple vision sensors

In this paper, we describe the monitoring of human activity in an indoor en vironment through the use of multiple vision sensors. The system described in this paper is made up of three cameras. Two of these cameras are active and are part of a binocular system. They operate either as a set of three s tatic cameras or as a set of one fixed camera and an active binocular visio n system. The human activity is monitored by extracting several parameters that are useful for their classification. The system enables the creation o f a record based on the type of activity. These logs can be selectively acc essed and provide images of the humans in specific areas. (C) 2001 Elsevier Science B.V. All rights reserved.