基于小波变换融合图像的全局运动估计新模型  

The Novel Model of Global Motion Estimation Based on Wavelet Transform Image Fusion

作　　者：陈艳浩[1] 赵胜杰[2] 马媛媛[2] 焦普 

机构地区：[1]河南师范大学网络中心,河南新乡453007 [2]河南师范大学计算机与信息工程学院,河南新乡453007

出　　处：《河南师范大学学报（自然科学版）》2014年第5期155-158,共4页Journal of Henan Normal University(Natural Science Edition)

基　　金：河南省信息技术教育研究项目(ITE12071)

摘　　要：像素运动的视频序列中全局运动方式占较大比重,这是由于摄像设备的移动所造成的.金字塔算法因为其工作原理会导致图像显示不稳定,小波变换对相关联的图像能进行有效的处理以克服图像融合技术金字塔算法的不足.提出了基于小波变换融合图像的全局运动估计的的新型模态,基于小波变换的多尺度结构可以通过融合不同图像的特征来解决.实验结果证明了该算法的有效性.Global motion means occupies a large proportion of the pixel motion,and it is caused by the movement of the camera in a video sequence. Key issues of Image Stabilization Algorithm are to obtain accurate global motion parameters byglohal motion estimation. Wavelet transforms to remove the correlation between the two adjacent sub-Image scales,so not only able to overcome the instability brought pyramid algorithm based on wavelet transform Image fusion technology. The paper puts forward the novel model of global motion estimation based on wavelet transform image fusion. Based on wavelet transform multiscale can be solved by Image characteristics of different image fusion brought difficulty. The experimental results indicate the effectiveness of the proposed algorithm.

关 键 词：全局运动估计 小波变换 图像融合 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]