Research of Sub-Pixel Image Registration Based on Local-Phase Correlation
A fast and effective sub-pixel image registration algorithm based on image local-phase correlation is presented. Using phase assessment methods, initial shift is performed by calculations. And the maximum cross-correlation value is obtained according to f
- PDF / 468,155 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 80 Downloads / 151 Views
Research of Sub-Pixel Image Registration Based on Local-Phase Correlation Shiwen Li and Xiaoxiao Liang
Abstract A fast and effective sub-pixel image registration algorithm based on image local-phase correlation is presented. Using phase assessment methods, initial shift is performed by calculations. And the maximum cross-correlation value is obtained according to fast Fourier transform (FFT) of the local power spectrum after upsampling on the local area of the image. The experimental results show that the algorithm exhibits high running speed and it can work on common computers; moreover, it can restrain noise well. Keywords Local-phase correlation • Fast fourier transform • Sub-pixel image registration • Image processing
64.1 Introduction Image registration is the process of overlaying images (two or more) of the same scene taken at different times, from different viewpoints and/or by different sensors [1]. Image registration is one of the most important steps in image processing, machine vision, and medical imaging. It is the basis of multi-source image analysis, and it has wide applications in the fields of remote sensing, motion estimation, computer vision, medical imaging, image fusion, and splicing as well as image enhancement and restoration [2]. In some cases, there is small portion of image needed to process or recognize. In dealing with remote sensing measurements, medical image recognition, image recognition, etc., the image is difficult S. Li (*) School of Engineering and Technology, Panzhihua University, Panzhihua 617000, China e-mail: [email protected] X. Liang School of Computer Science, Sichuan University of Science and Engineering, Zigong 643000, China e-mail: [email protected]
Z. Zhong (ed.), Proceedings of the International Conference on Information Engineering and Applications (IEA) 2012, Lecture Notes in Electrical Engineering 217, DOI: 10.1007/978-1-4471-4850-0_64, © Springer-Verlag London 2013
507
508
S. Li and X. Liang
to meet people’s requirements as a result of the complex conditions of image noise and movement of the imaging object [3]. Therefore, in order to get more efficient, faster image registration as well as able to been used in the harsh environment, people have undertaken extensive researches and put forward a variety of image registration algorithms [4]. In general, its applications can be divided into three main groups according to the manner of the image acquisition. Spatial correlation, the disadvantage of such method is that it cannot be used for the images of different resolutions and features. Transform-based registration method: it has been applied to the image of the same nature and not for heterogeneous image. Feature-based registration method: in such methods, we need to firstly find feature points which are appropriate for the given task and then the two images are registrated according to the feature points [5]. If we use the traditional method for data processing registration, such as fast Fourier transform (FFT), the operator computation would be ver
Data Loading...
