一种联合数据常量约束的分数阶光流算法  

A Fractional-order Optical Flow Model Based on Combined Data Constant Constraint

作　　者：李建[1] 汪兆栋[1] 

机构地区：[1]景德镇学院机械电子工程系,景德镇333000

出　　处：《科学技术与工程》2018年第6期299-303,共5页Science Technology and Engineering

摘　　要：针对传统HS(Horn&Schunck)变化光流模型难以应对图像中全局或者局部光照变化、非刚体运动、大位移运动、局部区域内存在多运动目标及难以保留边缘不连续性等实际应用问题,提出一种联合数据约束的分数阶光流模型;该模型分别对传统HS光流模型的数据项及平滑项进行了改进。数据项中,除了包含亮度常量约束外,同时还包含梯度常量约束(应对全局或者局部光照变化)、海森常量约束(应对大位移运动)、拉普拉斯常量约束(应对非刚体运动)。平滑项中,把传统HS的整数阶平滑约束方程改进为分数阶平滑约束方程,以此应对局部区域内的多运动目标问题及保留光流场运动边缘的不连续性。算法的性能体现在能应对各种对亮度常量约束的违反的实际应用;且能获得高准确率的光流场。大量试验验证了算法的优越性。In order to extent the original HS( Horn & Schunck) optical model to more complex environment,such as global or local illumination change,non-rigid motion,large displacement motion,multiple motions in local region and can't preserve the discontinuity of the edge of the optical flow field. A variational model for optical flow computation was proposed based on non-linearised and higher order constancy assumptions. Besides the common grey value constancy assumption,also gradient constancy,as well as the constancy of the Hessian and the Laplacian are included in the data term of the original HS model,meanwhile a fractional-order smoothness constraint equation was used to replace the integer order of smoothness constraint of the original HS model. The superior performance of the proposed method shows up by significantly smaller estimation errors when compared to previous techniques. Extensive experiments demonstrate the superiority of our algorithm.

关 键 词：图像序列分析 光流法 分数阶 拉普拉斯常量约束 海森常量约束 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]