A Novel Registration Method for High Resolution Remote Sensing Images Based on JSEG and NMI

Chao Wang1, 2, Aiye Shi1, 2, Xin Wang1, Fengchen Huang1 and Hui Liu1

  1. College of Computer and Information Engineering, Hohai University
    211100 Nanjing, China
    lxyliuhui@163.com
  2. Engineering Research Center of Sensing and Computing, Hohai University
    211100 Nanjing, China

Abstract

When traditional multi-scale analysis tools are applied to high resolution remote sensing image registration, difficulties and limitations are common in selection of directional sub-bands and distribution optimization of control point pairs etc. Aiming at this issue, a novel registration method based on JSEG and NMI is proposed in this paper. It is the method that incorporates the multi-scale segmentation method (JSEG) into image registration for the first time and proposes an adaptive feature point extraction method on the basis of blocking strategy. Then, NMI is adopted to obtain a set of control point pairs. Finally, the image registration is realized by virtue of Delaunay triangle local transform mapping functions. In accordance with experiments on high resolution remote sensing images collected by different sensors, it is found that the method can not only extract feature points accurately but also ensure reasonable distribution of control point pairs. Meanwhile, compared with traditional multi-scale tools-based methods, the method has relatively high accuracy and robustness.

Key words

Image registration; High resolution remote sensing images; JSEG; NMI

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS140505089W

Publication information

Volume 12, Issue 1 (January 2015)
Year of Publication: 2015
ISSN: 1820-0214 (Print) 2406-1018 (Online)
Publisher: ComSIS Consortium

Full text

DownloadAvailable in PDF
Portable Document Format

How to cite

Wang, C., Shi, A., Wang, X., Huang, F., Liu, H.: A Novel Registration Method for High Resolution Remote Sensing Images Based on JSEG and NMI. Computer Science and Information Systems, Vol. 12, No. 1, 289–306. (2015)