Border Crossing Detection and Tracking through Localized Image Processing
Aaron B. Greenblatt, Karen A. Panetta, Sos С. Agaian · 2008
The United States' borders are at risk and, due to their vast size, cannot possibly be monitored in their entirety by people at all hours of the day. Automated detection systems for border intrusion exist, but many of them require expensive electronics and lengthy installation procedures. This paper introduces a novel image processing algorithm that detects and tracks moving targets and, when coupled with infrared sensors, can detect temperature changes in the earth that may indicate paths being constructed for border-crossing under the surface. Intensive computer simulations show that the algorithm can also be adapted to automated vehicle control. The proposed algorithm can be implemented with commercial, off-the-shelf desktop computers and cameras, allowing for rapid deployment at low cost. Finally, use of Internet-enabled computers would allow a central command post to monitor the borders at all times and the algorithm could be combined with a simple system that automatically notifies command when a breach of the border occurs.