实时红外亚成像跟踪

Real-time Infrared Quasi-image Tracking

研究了玫瑰线扫描ＰＣ仿真的精确实现．研究了适合低对比度、空域模糊亚图像的子块二值迭代分割及空域模糊分割识别算法．提出了两种目标初始大概率检测策略，一是初始多帧配准识别，一是利用复合制导对指定区域中心进行扫描成像，通过红外亚图像序列运动补偿实时锁定跟踪，获得了高的跟踪准确度．最后对实验系统构成进行了说明．实验表明，本文提出的识别跟踪算法是可行的，能在ＤＳＰ实验系统中实时实现．

Precise synchronous scan is implemented with PC simulation in the paper.Iteraive subblock bilevel partition and spatial fuzzy segmentation are studied for low contrast and fussy infrared quasi image Two methods are investigated to improve original recognition rate,one is original recognition,and the other is imaging in appointed area by other combined guiding. Infrared quasi-image series with movement compensation and locking is adopted to get higher tracking probalility Finally experimental system is built. The experiment reveals that the recognition and tracking is feasile and they can be imlemented in real time in experimental system.