Signal Processing for a Southern Border Surveillance System

Successfully monitoring vast areas of landscape is crucial for maintaining security across the United States border. We envision the solution to this problem to consist of a search sensor, such as a radar, that cues an EO/IR sensor which provides recognition and tracking capabilities. In this Phase I effort, we propose to focus on the signal processing required to support the EO/IR recognition and tracking function. We propose a software solution that detects and tracks moving and stationary targets that may be closely separated. Our signal processing solution includes frame registration, frame differencing, target segmentation, target tracking, and target feature extraction. Such techniques have already established high target detection performance while maintaining low numbers of false alarms. We will address requirements for preventing intruders from crossing the Southern border and the considerations for retrieving illegal entrants that may evade border security. Further, we show how real-time processing capability can be achieved through efficient candidate target segmentation, followed by a hierarchical classification structure. Throughout, we demonstrate how storage and bandwidth requirements can be minimized, to enable deployment of the complete system in a variety of resource-constrained environments, leading to reduced operator workload and improved surveillance effectiveness.