"Smart Sensing and Tracking with Video and Mote Sensor Collaboration"
Sadiye Guler, Timothy Cole
IEEE International Conference--Technologies for Homeland Security, May 11-12, 2009
This paper describes research, design, and development of a feasibility prototype for video and mote sensor collaboration for efficient detection of security threats and propagation of alarms within a web-based sensor network environment. Our focus for this Smart Sensing and Video Understanding (SSVU) system has been three-fold: (1) to augment and cue the video sensors with cheaper auxiliary sensor motes and integrate these two sensor types within a multiple sensor platform to test and validate the prototype, (2) develop a prototype Smart Video Node for the Front-End Sensing Module which is suitable for operation in a smart sensor network environment, and (3) create device algorithms for processing of video data for detecting prescribed and/or anomalous events and security threats.
While the sensor networks may include several different sensor modalities operating over a wide range of data and application domains, our focus here is on the perimeter intrusion threats for applications involving large areas such as borders. For the protection of extended perimeters, the deployment of video cameras to cover the entire perimeter could be prohibitively expensive; therefore supporting video sensors with inexpensive and easily deployable unattended sensors provides an efficient solution. These auxiliary sensors can operate within a self organizing network to detect activity and can cue sophisticated video sensors to the position for a closer examination and evaluation of the potential threats.
* This work has been partially supported by a Department of Homeland Security Science and Technology Directorate Small Business Innovation Research award. The ideas and results presented here belong to the authors and do not represent that of DHS or United States government.