The Method of Infrared Point Target Detection and Tracking Based on DSP+FPGA

摘要

Accounting for the problems of dynamic infrared point targets are difficult to detect accurately and to track in complex background, the single-frame image background suppression by the high-pass filtering and binary processing. Then it uses unscented Kalman filter (UKF) method to get the predicted target trajectory and achieves a precise target tracking. According to the proposed detection and tracking algorithm of motion infrared point target, this paper designs and implements a system of high speed real-time image processing platform based on DSP+FPGA structure. Test results show that when the target is in the interference condition, tracking can still be stabilized, its processing time is about 9.2ms per frame to meet real-time requirements.