矿井移动载体摄像的电子稳像算法  

An electronic image stabilization algorithm of cameras for underground moving carriers

作　　者：程德强[1] 赵国[1] 范一武[1] 安铎[1] 钱建生[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221116

出　　处：《中国矿业大学学报》2012年第5期805-810,共6页Journal of China University of Mining & Technology

基　　金：国家高技术研究发展计划(863)项目(2008AA062200);高等学校博士学科点专项科研基金项目(20110095120007);徐州市科技项目(XF11C110;XX10A001)

摘　　要：针对矿井移动载体摄像造成的视频抖动,提出一种电子稳像算法,分离移动载体摄像中的主动扫描运动矢量和随机抖动矢量,实现视频序列的稳像补偿.算法首先利用平移运动模型建立图像运动矢量估计方程,以视频序列的角点特征建立帧间映射,求解其最小二乘解得到帧间图像全局运动矢量.然后针对移动载体摄像抖动的不确定性,构建粒子状态预测方程,采用二维欧氏距离定义观测值和观测概率密度,求解运动补偿矢量.通过算法有效性和稳像准确度实验,结果表明:该稳像算法可有效保留摄像机主动扫描运动、去除随机抖动,提高了输出视频的图像质量,稳像后峰值信噪比(PSNR)的平均值比稳像前提高了25.38DB.The quality of a video is greatly influenced by the irregular vibrations from the cam- era for underground moving carriers. In this work, an electronic image stabilization algorithm is proposed to improve the stability and fidelity of video sequence from cameras for moving car- riers by removing the inter-frame vibration between video sequences. Firstly, the equations set to solve the vectors of motion are established with the moving transformation model. By ex- tracting the Harris corners, the mapping relationship between the candidate frame and the cur- rent frame is built based on the calculated results of the equations set as global motion vectors. Then, according to the uncertain vibrating positions of moving carriers, the motion offset vec- tors are computed by establishing the state--forecasting equations for the particles. The obser- vation value and its probability density function are defined by the 2D Euclidean distance. The experiments of validity and fidelity about algorithm show that the proposed algorithm can sepa- rate the active scanning motion vectors form random vibrations of camera, which achieves to improve the quality of output video sequence. The average peak signal to noise rate （PSNR） was increased to 25.38 DB after the vibrating video sequence was stabilized.

关 键 词：电子稳像 运动估计 粒子滤波 

分 类 号：TN391[电子电信—物理电子学]