摘要
根据铜转炉内壁蚀损趋势分析的需要,提出了一种基于等温线轮廓特征的红外系列图像配准算法。使用了质心距离函数和傅里叶描述寻找相似度最高的等温线轮廓,在相似轮廓线对的点集合上建立对应关系,经匹配将图像配准问题转换为标准的参数矩阵最小二乘估计问题,由仿射变换建立图像的配准空间。测试结果表明,算法具有简单、运算速度快的特点,能满足分析精度要求。
For the tendency analysis of copper converter inner-lining wear, a new method of infrared sequential nnages registration based on the isotherm contour features was proposed. The most similar isotherm contour was sought by centroid distance function and Fourier descriptor. The parallelism point set was given on the similar isotherm contour, and the images registration space was built by projective transformation. The image registration problem was converted into the least square estimation problem of standard parameter matrix. The test result shows that the isotherm contour matching and images registration are found rapidly and simply by the method, which may attain the analytical precision of smelting plant.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2008年第11期96-100,121,共6页
Opto-Electronic Engineering
基金
国家自然科学基金资助项目(60764002)
江西省自然科学基金资助项目(0650041)
关键词
图像配准
红外图像
等温线
炉衬蚀损
最小二乘法
image registration
infrared image
isotherm
inner-lining wear
least square method
作者简介
刘飞飞（1962-），男（汉族），江西赣州人，教授，博士，主要从事红外检测、图像测量、光机电一体化研究。E-mail：gzlff@126．com
