高速水面艇视觉系统电子稳像算法  

Electronic image stabilization algorithm for high speed surface vehicle vision system

作　　者：马忠丽[1] 李慧凤[1] 文杰[1] 梁秀梅[1] 

机构地区：[1]哈尔滨工程大学自动化学院

出　　处：《计算机应用研究》2014年第2期633-636,共4页Application Research of Computers

基　　金：国家自然科学青年基金资助项目(51109047);国家留学基金委留学基金资助项目(2011307358);黑龙江省博士后基金资助项目(Ibhq10140)

摘　　要：针对高速水面艇视觉系统在采集视频过程中,由于高速运行、水流运动和风力影响等因素造成的视频图像抖动问题,根据高速水面艇视频图像运动特点,例如同时含有平移、旋转和变焦运动等,采用尺度不变特征变换算法提取视频图像中的特征点,利用仿射模型求解运动参数,运用Kalman滤波对视频图像中的正常扫描进行滤波,最后用相邻帧补偿法对每帧图像进行补偿,实现高速水面艇的视频图像稳像处理。算法用于高速水面遥控艇采集到的视频上进行对比验证分析,结果表明算法对高速水面艇视觉系统下的视频图像稳像处理快速、有效。Because of high-speed moving, flow movement, wind effects and other factors, the video captured by vision system of a high speed surface vehicle always contained blurring, according to moving characteristics of video images of a high-speed surface vehicle, such as the translation, rotation, and zoom were included at the same time in an image, this paper used the scale-invariant feature transform algorithm to extract feature points of the video image sequence, and used affine model to solve the motion parameters, and then used Kalman filter to deal with the normal scan of the images sequence. Lastly, it adopted adjacent frames compensation method to compensate each frame of the video image. Using proposed algorithm deals with a sec- tion of video with blurring from a high-speed remote-control surface vehicle, the experimental results show that the algorithm is fast and effective to solve video image stabilization for a high-speed surface vehicle vision system.

关 键 词：高速水面艇 电子稳像 尺度不变特征 仿射模型 KALMAN滤波 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]