红外弱小目标检测跟踪与实现  

Method of small target detection and tracking from infrared image and its realization

作　　者：李建勋[1] 张世俊[1] 陈非[1] 刘刚[1] 刘广智[1] 敬忠良[1] 

机构地区：[1]上海交通大学电子信息学院信息与控制研究所,上海200030

出　　处：《系统工程与电子技术》2006年第1期11-14,118,共5页Systems Engineering and Electronics

基　　金：国家自然科学基金(60304007);航空科学基金(03F57003);国防科技重点实验室预研基金(51483020203JW0310;51476010604JW0303);上海交通大学青年教师基金(51483020203JW0310)联合资助课题

摘　　要：针对红外序列图像中弱小目标的检测问题,提出了一种基于旋转均值滤波的弱小目标检测方法,同时利用数据关联进行多帧积累实现序列图像跟踪。基于Hunt Engineering的HEPC9、双DSPC6201和双FPGA搭建了一套实时红外图像处理系统,并将上述检测、跟踪算法成功移植于DSPC6201硬件处理器。实测数据的实时处理证明,上述方法可实时和有效地检测和跟踪低信噪比(≈2)复杂序列图像,系统处理能力为50帧/s,320*240/帧,14位/像素(双6XDSP)。A novel detection algorithm based on rotating mean filter is proposed for the detection of small moving target from infrared image. The data association is made to accumulate the sequences of infrared image. A real-time image processing system with Hunt Engineering HEPC9, double-DSPC6201, and double-FPGA is built, and the above detection and tracking algorithm is successfully embedded into the hardware processor. Processing results with the real image sequences show the proposed method can detect and track a small moving target from low SNR（≈2） infrared image sequences in real time and effectively, the processing capacity is 50 frames/s, 320 ＊ 240/frame, 14 bits/pixel （double 6XDSP）.

关 键 词：旋转均值滤波 数据关联 双DSPC6201硬件处理器 

分 类 号：TN911.73[电子电信—通信与信息系统]