基于椭球面模型的鱼眼图像校正方法  

A fisheye image correction method based on ellipsoid model

作　　者：杨中伟[1] 杨鸣[1] 

机构地区：[1]宁波大学信息科学与工程学院,浙江宁波315211

出　　处：《光学仪器》2015年第2期127-131,共5页Optical Instruments

基　　金：宁波市重大科技攻关项目(2010B10007)

摘　　要：鱼眼镜头的大视角、短焦距等特点,使鱼眼图像校正算法成为目前图像研究的热点。提出一种椭球面校正模型,通过边缘检测算法获取特征点,运用伪特征点剔除法得到正确的特征点,利用这些真特征点得出校正模型的参数,并以网格图像为例进行研究。实验结果表明,特征点检测准确,得到的模型对网格图像的校正效果好,验证了该校正模型的准确性和有效性。Fisheye lens have the characteristics of short focus and wide-angle,which makes the fisheye image correction algorithms become hot topics in the field of image research currently.An ellipsoid correction model is put forward.All feature points are obtained by using edge detection algorithm,and the correct feature points are also obtained by using the pseudofeature-point elimination method.The correction model parameters are obtained with the correct feature points.The grid image is taken as an example to conduct the research.As a result,the detection of the feature points is accurate and the correction effect is good.It is proved that this model is accurate and efficient.

关 键 词：鱼眼镜头 椭球面模型 伪特征点剔除 畸变校正 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]