舰船图像序列的电子稳像技术  

Electronic image stabilization of ship image sequence

作　　者：冯驰[1] 覃春花[1] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《应用科技》2010年第1期15-18,共4页Applied Science and Technology

摘　　要：针对船载摄像系统获取的图像序列抖动的问题,提出一种基于信息融合的电子稳像方法,对因风浪作用产生晃动的视频图像序列进行稳定处理,使输出的视频图像序列稳定而清晰.基于信息融合的电子稳像方法是指在运动估计过程中,对于传感器获取的运动矢量与软件算法获取的运动矢量进行信息融合,得到更为准确和有效的运动矢量,并对视频图像序列进行补偿的一种方法.给出采用软件方法、硬件方法和信息融合3种方法获取运动矢量,进行补偿得到的结果.与参考图像进行比较分析,结果表明信息融合的方法可以使稳定精度大大提高,更适合人眼观察.In this paper we introduce an stabilization algorithm based on Information Fusion on the characteristics of image sequence acquired in a ship. The shaking video image sequence under the influence of wind and waves is stabilized treatment, so that the output of the video image sequence is stable and clear. And electronic image stabilization based on information fusion is defined as motion estimation with information fusion. Motion vectors are obtained from information fusion between for the sensors and software algorithms, so that the vectors are more accurate and effective. Then compensate the video image sequence and export the stable and clear video image sequence. The paper describes the three methods in the use of software, hardware methods and information fusion access to motion vectors, then compensate them. The compensatory results and a comparative analysis of reference images are given in the article in the end. The results show that information fusion method can greatly improve the accuracy and stability, and is more suitable for observation of the human eye.

关 键 词：电子稳像 运动估计 硬件方法 软件方法 信息融合 

分 类 号：TP317[自动化与计算机技术—计算机软件与理论]