基于光照辐射模型的光流场估计  

Optical Flow Field Estimation Based on Illumination Radiation Model

作　　者：廖彬[1,2] 杜明辉[1] 胡金龙[3] 

机构地区：[1]华南理工大学电子与信息学院,广东广州510640 [2]华南农业大学理学院,广东广州510642 [3]华南理工大学信息网络工程研究中心,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2011年第5期44-48,54,共6页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家自然科学基金重点项目(U0735004);国家科技支撑计划项目(2008BAH37B08)

摘　　要：为完成多种环境下针对不同物体的运动检测,提出了基于光照辐射的光流场模型.首先,经由通用动态图像模型在克利福德代数域上的分析,解释并修正了通用模型的参数;然后,利用克利福德代数域上的推导结果与后期实验数据,结合运动视觉产生原理与光照辐射理论,提出了基于光照辐射的光流场模型,并利用梯度场提高其准确性;最后,对此模型给出总变分法.对比实验结果显示,文中提出的模型具通用性,并能在不同光照条件下获得准确且连续一致的光流场.Proposed in this paper is an illumination radiation-based optical flow field model for the motion detection of various objects in different environments. In the investigation, first, the parameters of the general dynamic image model are explained and amended in the Clifford algebra region. Then, according to the results deduced in the Clifford algebra region and obtained from experiments, a new optical flow field model is proposed based on the principle of movement vision-generating principle and the illumination radiation theory, to which a gradient field is applied for the purpose of improving the accuracy. Finally, a total variation algorithm is presented for the proposed model. Experimental results show that the proposed model can achieve accurate, continuous and consistent optical flow fields for the purpose of being generally used in different illumination conditions.

关 键 词：光流场模型 克利福德代数 光照辐射 总变分法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]