A scalable architecture system for automatic target recognition

Automatic target recognition (ATR) is a very difficult problem due to the variety of conditions under which an ATR system may be required to operate. Because the number of operations required to execute a particular ATR algorithm can vary greatly from one scenario to another; a fixed hardware and software architecture will usually not be able to execute a given ATR algorithm in all required scenarios within some given real-time constraints. A solution to this problem is to use a scalable architecture. The hardware and software of such an architecture can easily be scaled to meet the processing requirements of a particular scenario. This paper describes a scalable architecture system which we have developed that implements a real-time ATR algorithm.<
>