Automated Cueing Using Motion Detection and Extraction

To address SOCOM needs for advanced man portable surveillance technologies, Physical Optics Corporation (POC) proposes to develop a new system for Automated Cueing using Motion detection and Extraction (ACME). ACME is structured around a new hybrid motion detection and object recognition software kernel based on feed-forward neural networks, and state-of-the-art statistical algorithms, capable of identifying objects, and of detecting motion at ranges of 1000 m in daylight and 500 m at night. The ACME system, interfacing imagers and processing hardware, has intelligent and versatile surveillance capability with visual/audible alerts and wired/wireless network interfaces. ACME interfaces with existing imaging systems and data processing hardware through a universal video interface compatible with USB, FireWire, and any video capture-capable device, and thus requires no additional hardware for operation. Its high performance video processing software eliminates the need to alter or manipulate the imager, which will have accelerated video processor capabilities including filtering, geometrical transformation, digital zoom, and fast video display. In Phase I the video converter box and each of the core software modules for ACME will be developed and demonstrated. In Phase II the ACME software modules will be combined into an ACME system suitable for immediate transition to Phase III markets.