摘要
本文应用线性和非线性目标函数计算图象序列中物体的速度(光流)场.用松弛法计算速度场的同时,这些目标函数的参量也被估计出来,目标函数既提供了动态景物的描述,又可作为平滑性的约束.计算速度场时利用测量边界或区域界的垂直速度分量的迭代调整算法克服了“孔径问题”的困难.用分面线性和二次型目标函数对物体的三维运动进行了分析和描述.
Image velocity(optical flow)fields are Computed and interpretedfrom an image sequence by incorporating the knowledge of object kinematics.Linear and nonlinear objective functions are defined.The function parametersare estimated simultaneously with the computation of the velocity field byrelaxation.The objective functions provide an interpretation of the dynamicscenees,and at the same time serve as smoothness constraints.Thecomputation is initially based on measured perpendicular components ofcontours or regional boundaries.The difficulty of the “aperture problem”isalleviated by introducing iterative adjustment of measurement directions.Theapproach lends itself to the analysis of 3D motion of a solid object,usingpiecewise objective functions.As an example,the velocity fields generated bythe image sequence of an automobile in motion were consistent with visualperception.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
1989年第5期73-82,共10页
Journal of Shanghai Jiaotong University
关键词
目标函数
图象序列
光流
速度场
Image sequence
optical flow
objective function
motion estimation
