Registration of infrared and visual images based on phase grouping and mutual information of gradient orientation

This paper presents a novel infrared and visual image registration method based on phase grouping and mutual information of gradient orientation. The method is specially designed for infrared image navigation, which is different from familiar multi-sensor image registration methods in the field of remote sensing. The central idea is to firstly extract common salient structural features from visual and infrared images through phase grouping, then registering infrared image to visual image and estimating the exterior parameters of the infrared camera. Two subjects are involved in this reports: (1) In order to estimate image gradient orientation accurately, a new method based on Leguerre-Gauss filter is presented. Then the image are segmented by grouping of pixels based on their gradient orientations and ling support regions are extracted as common salient structural features from infrared and visual images of the same ground scene. (2)In order for registering infrared and visual image, coordinate systems are constructed, coordinate transformations are formularized, and the new similarity measures based on orientation mutual information is presented. Quantitative evaluations on real and simulated image data reviews that the proposed method can provide registration results with improved robustness and accuracy.

[1]  Max A. Viergever,et al.  Image registration by maximization of combined mutual information and gradient information , 2000, IEEE Transactions on Medical Imaging.

[2]  Liu Jian-ye.  Research on the Fast Scene Matching Algorithm in the Inertial Integrated Navigation System , 2009 .

[3]  Zhuo Pan,et al.  A High Time-Efficiency Multi-modal Images Matching Scheme for Navigation: A High Time-Efficiency Multi-modal Images Matching Scheme for Navigation , 2011 .

[4]  Carlo Tomasi,et al.  Image Similarity Using Mutual Information of Regions , 2004, ECCV.

[5]  Amir Averbuch,et al.  Multisensor image registration via implicit similarity , 2006, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Daniel Rueckert,et al.  Non-rigid registration using higher-order mutual information , 2000, Medical Imaging.

[7]  Cheng-Chang Lu,et al.  Multi-modality Image Registration Using Mutual Information Based on Gradient Vector Flow , 2006, 18th International Conference on Pattern Recognition (ICPR'06).

[8]  Guy Marchal,et al.  Multimodality image registration by maximization of mutual information , 1997, IEEE Transactions on Medical Imaging.

[9]  Peter Kovesi,et al.  Image Features from Phase Congruency , 1995 .

[10]  Mark Holden,et al.  Multi-channel mutual information using scale space, in Medical Image Computing and Computer-Assisted Intervention , 2004 .

[11]  Juan Su,et al.  A Multi-sensor Image Registration Algorithm Based on Structure Feature Edges: A Multi-sensor Image Registration Algorithm Based on Structure Feature Edges , 2009 .

[12]  Chen Wu.  A Rapid Image Matching Approach for Auxiliary Navigation , 2009 .

[13]  Atr Key,et al.  An Optimal Algorithm for IR/Visual Image Registration Based on Main-Line-Pairs , 2007 .

[14]  Jong Beom Ra,et al.  Multi-sensor image registration based on intensity and edge orientation information , 2008, Pattern Recognit..

[15]  Jong Beom Ra,et al.  Robust CCD and IR Image Registration Using Gradient-Based Statistical Information , 2010, IEEE Signal Processing Letters.

[16]  Lin Xing.  A Multi-sensor Image Registration Algorithm Based on Structure Feature Edges , 2009 .

[17]  B. S. Manjunath,et al.  A contour-based approach to multisensor image registration , 1995, IEEE Trans. Image Process..

[18]  Chen Wu-fan.  Medical Image Registration Based on Principal Phase Congruency , 2008 .

[19]  Allen R. Hanson,et al.  Extracting Straight Lines , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[20]  P. Anandan,et al.  Robust multi-sensor image alignment , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).