基于红外多目标图像序列的自动判读技术  

Automatic Interpretation Algorithm Based on Infrared Multi-target Sequence

作　　者：李铫[1,2] 刘文[1,2] 刘朝晖[1] 

机构地区：[1]中国科学院西安光学精密机械研究所,陕西西安710119 [2]中国科学院研究生院,北京100039

出　　处：《现代电子技术》2009年第2期154-157,共4页Modern Electronics Technique

摘　　要：提出了一种红外多目标图像序列的自动判读方法,先着重阐述一种目标序列段自动变步长搜索方法精确定位有用目标序列段,然后对图像二值化分割批处理方法以及判读脱靶量计算等关键性技术进行说明,最后输出准实时脱靶量;探讨目标搜索中步长和信号量阈值对搜索时长以及精度的影响。实验结果表明,对于低SNR的红外弱小多目标图像序列,该判读方法能缩短判读时间,提高判读的准确性。A method of automatic interpretation for infrared multi - target image sequence is proposed. An automatic variable step size searching method of target sequence is applied to locate useful target segments precisionly. Then, key techniques such as image group binaryzation method and interpretation target - missing quantity calculation are discribed. Finally, interpretation results are calculated and exported in real- time. In addition,it discusses the influence of searching time caused by step size and signal measuring threshold in target searching. The results show that the time and precision and precision of interpretation are greatly improved for the image sequence with low SNR.

关 键 词：自动判读 自动变步长搜索 脱靶量 图像处理 

分 类 号：TP753[自动化与计算机技术—检测技术与自动化装置]