摘要
基于非线性Ｋａｌｍａｎ滤波器的长序列立体图象运动估计黄浴，袁保宗，彭光凤（北方交通大学信息科学研究所，电气工程系，北京１０００４４）摘要：讨论一种基于运动模型的、从长序列含噪立体图象中估计目标运动参数的方法。目标的运动模型具有恒进动的旋转分量和恒加速...
In this paper, a kinematic model-based approach of estimating the object motion parameters from a long sequence of noisy stereo images is discussed. The object's modeled motion has constant precession in the rotational components and constant acceleration in the translational components, and an iterative Kalman filter with damped factors is designed to estimate the model's parameters. To alleviate divergence of the nonlinear filter, a linear algorithm of generating initial state guess is proposed. Eventually,results of computer simulation are given.
出处
《北方交通大学学报》
CSCD
北大核心
1996年第2期124-130,共7页
Journal of Northern Jiaotong University
基金
国家"863"高科技资助项目
关键词
长序列立体图象
运动模型
卡尔曼滤波器
非线性
s: long sequence of stereo images, kinematic model, estimation of motion parameter
