Remote sensing image registration method based on anisotropic gradient dimension space

Abstract

The invention discloses a remote sensing image registration method based on anisotropic gradient dimension space, which mainly solves the problem of relatively low correct matching rate under the condition of relatively great brightness nonlinear change of the remote sensing images. The implementing steps of the remote sensing image registration method based on anisotropic gradient dimension space are as follows: (1) inputting remote sensing image pairs; (2) constructing dimension space of anisotropic diffusion; (3) calculating a gradient amplitude image; (4) detecting feature points; (5) generating a main direction of the feature points; (6) generating a descriptor of each feature point; (7) matching the feature points; (8) deleting wrongly matched feature point pairs; and (9) registering a reference image and a to-be-registered image. As feature point detection, feature point main direction generation and feature point descriptor generation are carried out on the gradient amplitude image in the anisotropic dimension space, the situation of relatively great brightness nonlinear change of the images can be dealt efficiently, and the remote sensing image registration method based on anisotropic gradient dimension space can be applied to complex multisource and multispectral remote sensing image registration.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a remote sensing image registration method based on anisotropic gradient scale space in the technical field of remote sensing image registration processing. The invention can be applied to registration of remote sensing images obtained at different times from multiple spectra and sources. Background technique [0002] Image registration refers to the process of geometrically correcting images of overlapping areas captured by the same or different sensors in the same scene from different perspectives at different time periods. Image registration technology is one of the image processing technologies developed rapidly in recent years. Image registration technology has been widely used in various fields, such as aerospace technology, image mosaic, geographic information system, image fusion, 3D reconstruction, object recognition and change detection, etc. At present...

AI Technical Summary

Problems solved by technology

However, the image registration method based on grayscale has the following disadvantages: ① it is sensitive to the change of image grayscale, especially the nonlinear illumination change, which greatly reduces the performance of the algorithm; ② the calculation complexity is high; ③ the target rotation, Sensitive to deformation and occlusion

The algorithm controls the quantity, quality and distribution of the feature points extracted in the scale space better, and improves the registration accuracy of some remote sensing images with local transformations, but the algorithm still has the disadvantage that it cannot accurately register Remote sensing image pairs with large nonlinear changes in image brightness

The shortcomings of this method are that since this method is based on mutual information, the computational complexity is high, and the registration accuracy is lower than the initial solution obtained by the sift algorithm. When there is a large nonlinear change in the image brightness In this case, the sift algorithm cannot obtain a good initial solution, and the mutual information method relying on the initial solution obtained by the sift algorithm cannot achieve accurate registration

Although this method can speed up the speed of image registration, the shortcomings of this method are that the SIFT method is still used in the stages of feature point generation, feature point main direction generation and feature point descriptor generation. The correct matching rate drops rapidly when the brightness nonlinearity changes greatly. At the same time, this method uses the traditional nearest neighbor and second nearest neighbor distance ratio matching criteria. When there are many repeated features in the remote sensing image, the number of correctly paired feature points is also rapid decline

Images