基于主成分分析的Fourier-Mellin变换在图像匹配中的作用  

Fourier-Mellin Transform Based on Principal Component Analysis in Image Matching

作　　者：熊文真[1] 甘露[1] 

机构地区：[1]信阳职业技术学院,河南信阳464000

出　　处：《柳州职业技术学院学报》2016年第2期45-49,共5页Journal of Liuzhou Vocational & Technical College

摘　　要：为了解决Fourier-Mellin描述子在图像匹配求解过程中计算量大、存在插值误差、旋转角度ψ和尺度缩放因子t不准确等问题。文章提出了一种新的算法,基于主成分分析的Fourier-Mellin描述子匹配算法,先利用主成分提取图像的重要特征,这样就把错综无规律的变量降低为较少的几个主要变量,然后对其重要特征进行匹配,这样能够减少计算量,实现图像的快速匹配,实验结果证明该算法的有效性。In order to solve the Fourier-Mellin descriptor in image matching which made the solving process of large amount of calculation, there are problems of interpolation error, rotation angle ψ and scaling factor t is not accurate. This paper presents a new algorithm which is an important feature of Fourier-Mellin descriptor based on principal component analysis of the matching algorithm. First, use of the main components of the image are extracted. Thus, the intricate irregular variable is reduced to a few main variables. Finally, matching its an important feature can reduce the amount of calculation, fast matching of image. The experimental results show that the effectiveness of the algorithm.

关 键 词：主成分分析 图像匹配 Fourier-Mellin描述子 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]