矿井车载摄像系统中电子稳像算法的研究  

Electronic Image Stabilization of Camera System for Underground Vehicle Carries

作　　者：程德强[1] 郭政[1] 郑珍[1] 姜海龙[1] 刘洁[1] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221008

出　　处：《电视技术》2015年第15期19-23,共5页Video Engineering

基　　金：国家自然科学基金委员会与神华集团有限责任公司联合资助项目(U1261105);江苏省产学研前瞻性联合研究项目(BY2013019-6)

摘　　要：为克服矿井车载摄像系统在视频采集过程中的图像抖动问题,提出了一种基于MIC-Harris角点特征的自适应运动补偿电子稳像算法。首先,提出MIC-Harris角点检测算法和由粗到精的匹配策略,实现序列中特征点的快速精确提取。然后,根据矿井移动车载图像运动特点,构建运动参数估计模型,运用最小二乘法实现全局运动矢量的求解。最后,提出自适应相邻帧补偿方法,结合Kalman滤波实现补偿矢量的计算,在滤除高频随机抖动同时保留了车载摄像机低频主动运动,并克服了连续视频序列长时间稳像处理的帧间累积误差问题。实验仿真结果表明:提出的算法不仅融合了MIC运算速度快和Harris准确度高的优点,而且能够实现矿井车载摄像系统的长时间连续稳像处理,具有较好的稳像效果。An adaptive motion compensation electronic image stabilization based on MIC-Harris corner feature was proposed to solve tile jitter problem in the video acquired from underground-vehicle camera system. Firstly, the MIC-Harris corner detection algorithm and the matching strategy from coarse to precious are presented to extract feature points accurately in the sequence. Then,the motion parameter estimation model is built and the global motion vet, tots are computed by the least squares method ac- cording to the underground-vehicle motion characteristics. Finally, the adaptive adjacent fi＇ames compensation method is proposed to calculate compensation vectors combined with Kabnan filter. Not only the high frequency random jitter is fihered but also the low frequency active motion is retained, and the inter-frame cumulative error problem occurred by prohmged stabilization process of the continuous video sequences is also solved. The experiments show that the proposed algorithm in this paper can not only fuse the fast computing speed of MIC and high accuracy of Harris, but also achieve long continuous stabilization processing of underground-ve- hicle camera system. The proposed algorithm has a good stabilization effect.

关 键 词：矿井电子稳像 MIC-Harris 卡尔曼滤波 帧补偿 

分 类 号：TP319.4[自动化与计算机技术—计算机软件与理论]