一种基于方向平滑约束优化的多尺度光流算法  

A multi-scale optical flow algorithm based on oriented smoothness constraint optimization

作　　者：张泽旭[1] 崔平远[1] 崔祜涛[1] 

机构地区：[1]哈尔滨工业大学深空探测基础研究中心信息处理实验室,哈尔滨150001

出　　处：《高技术通讯》2007年第2期153-158,共6页Chinese High Technology Letters

基　　金：国家自然科学基金（60305003）和863计划（2005AA738041）资助项目.

摘　　要：提出了一种采用方向平滑约束的多尺度光流算法。该算法在多尺度的拉普拉斯金字塔框架下，利用差值平方和（SSD）区域匹配方法实现对搜索区域内的流速及其协方差矩阵的估计，然后利用方向平滑约束对估计的流速进行传播和优化，方向平滑约束可以有效地抑制对遮挡和运动不连续处的过度平滑。人工和真实图像序列的仿真实验都获得了满意的结果。A multi-scale optical flow algorithm was proposed, which uses oriented smoothness constraint to propagate and optimize the local and rough estimation of the optical flow. Under the frame of multi-scaled Iaplacian pyramid, the sum-ofsquared difference （SSD） matching technique estimates the flow velocity and the covariance matrix in searching area. Then the oriented smoothness constraint is employed to propagate and optimize the flow velocity estimated, which can suppress the over-smoothness across the occlusions area and motion discontinuities. The experiments obtained encouraging results using synthetic and real image sequences.

关 键 词：光流 多尺度 拉普拉斯金字塔 方向平滑约束 差值平方和 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]