Image registration method based on global information

Abstract

The invention discloses an image registration method based on global information, which relates to the image processing field and is used for improving the accuracy of image registration and enhancing the robustness of the registration method. The image registration method provided by the invention uses the global information of the image. For measuring the registration degree of the image, the definition and calculation of a maximum variance changing field of the image and the image registration measure based on an arithmetic-geometric inequality are provided. The technical scheme provided by the invention comprises the following steps of: calculating the maximum variance changing field of the image, calculating the registration measure between images based on the maximum variance changing field, improving the registration measure, optimizing the registration parameters, and executing registration on the images according to the registration parameters. The technical scheme provided by the embodiment of the invention is suitable for an image processing system which needs to perform registration on the images.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image registration method. Background technique [0002] In image processing and application, in order to obtain accurate and comprehensive information, it is sometimes necessary to fuse or detect differences of images of different modes or at different times of the same object. In these applications, image registration is a prerequisite. There are two types of image registration methods based on pixels and features. Among them, in pixel-based methods, mutual information is often used in image registration as a measure coefficient of image registration degree. The traditional mutual information measures the statistical correlation of image grayscale, only uses the grayscale information of the image, and does not consider the spatial information. Therefore, some image registration methods based on mutual information may have misalignment. Someone proposed the concept of m...

AI Technical Summary

Problems solved by technology

However, for images of different modes, due to the different gray characteristics, the maximum distance gradient field is difficult to accurately reflect the changes in the image

At the same time, computing gradients directly using grayscale makes such fields susceptible to noise

Images