A knowledge-based infrared search system in complicated IR background

This paper describes a knowledge-based infrared (IR) search system for the detection and classification of target formations in IR image sequences in complicated IR backgrounds. The system employs a fast fuzzy-based approach for detecting and recognizing targets, and utilizes some advanced technology such as spatial digital filtering, target acquisition by frame to frame, knowledge-based synthetic decision making, and also uses digital signal processors (DSPs) and field programmable gate arrays (FPGAs) to construct a real-time parallel processing architecture. The system is capable of automatic multi-target recognition in complicated IR backgrounds. The system processing rate reaches 30 frames per second, with a system response time of less than 1 second.