基于Zynq的空空导弹末端识别与定位系统研究

Research on Terminal Recognition and Location System of Airborne Missile Based on Zynq

针对目前空空导弹末端识别与定位速度低下的问题,提出一种基于显著性分析的红外小目标检测算法,结合SVM分类器,实现飞机红外目标的快速检测与识别。硬件利用Zynq-7000处理器,并基于Vivado HLS对显著性分析中的频域残差计算模块进行高层次硬件综合,构建硬件IP核,实现FPGA端的专用算法硬件加速。在Zybo平台上搭建嵌入式图像处理系统,以某大型运输机和某型战斗机为目标,实验结果表明,运输机和战斗机的识别率达到99%以上,定位速度达到每帧0.25 s。

In order to solve the problems of low speed of terminal recognition and location of airborne missile,an infrared small target detection algorithm based on saliency analysis is proposed,combined with SVM classifier,the fast detection and recognition of aircraft infrared target is realized.This paper gives full play to the advantages of the Zynq-7000 embedded processors which including ARM+FPGA.Based on Vivado HLS,the frequency domain residual algorithm is synthesized,and its hardware IP core is constructed to achieve hardware acceleration at the FPGA part,IP core is called by software algorithm of ARM to improve image processing speed.An embedded image processing system for airborne missile terminal is built on Zybo platform,and a transport aircraft and a fighter aircraft are used as experimental objects.The experimental results show that the recognition rate of transporter and fighter is over 99%and the positioning speed is 0.25 s per frame.