一种基于单个神经元的摄像机标定自适应算法  

An adaptive algorithm of camera calibration based on single neuron

作　　者：袁野[1] 欧宗瑛[1] 

机构地区：[1]大连理工大学机械工程学院CAD&CG研究所

出　　处：《大连理工大学学报》2005年第6期823-826,共4页Journal of Dalian University of Technology

摘　　要：传统的标定算法利用标准的参照物与图像点的对应约束关系来求取摄像机参数,其中非线性优化方法标定精度较高,但计算繁琐.为此提出一种基于单个自适应神经元的摄像机传统标定算法,应用一种结构简单、抗干扰能力很强的单个神经元自适应算法代替通常的非线性优化算法进行摄像机标定.实验结果表明该算法无需计算雅可比矩阵,且精度较高,简单可行.Conventional calibration algorithms are used to get the camera parameters by the relations between the standard calibration plane points and the according image points. Among these algorithms, nonlinear algorithms are precise, but very complex. So a conventional calibration algorithm based on single adaptive neuron is proposed, a simple and anti-disturbed single adaptive neuron is used to replace the nonlinear optimization algorithm. The experimental results show that compared with the former conventional optimization algorithm, this algorithm doesn＇t need the calculation of Jacobian matrix, and is precise, simple and feasible.

关 键 词：自适应神经元 摄像机标定 摄像机模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]