基于双向运动估计的残差分布式视频重构模型研究

摘要

针对遥感视频成像卫星星上编码端结构复杂、传输数据量大及传输带宽大的问题,同时为了保证遥感视频成像系统的高图像分辨率和低采样复杂度,根据视频图像的高时空冗余特点,在连续图像残差重构模型的基础上提出了基于双向运动估计的分布式视频残差重构模型及相关算法,并进行了数值仿真.仿真结果表明,模型能够在降低采样率的同时保证良好的重构效果,且算法简单,耗时短.%Some problems exist inremote sensing video imaging satellitessuch as the high complexity of coders,the large volume of transmitted data and the increased transfer bandwidth.To ensure the high resolution andreduce the complexity of remote sensing video imaging systems,this paper,according to the high time-space redundancy of video image,proposes a distributed residual reconstruction model and associated algorithm using bilateral motion estimation based on the technique of reconstructing frames of the video sequence using residual information.The simulation shows that the proposed model and algorithm ensure the low sampling rate and high reconstruction quality simultaneously,the algorithm is simple,andthe time needed is reduced.