SAR image registration method based on salient division sub-region pair

Abstract

The invention provides an SAR (Synthetic Aperture Radar) image registration method based on a salient division sub-region pair. The SAR image registration method comprises the following steps: 1) inputting SAR images of two time phases in the same region; 2) carrying out average filtering and histogram equalization on the two SAR images to suppress coherent spots; 3) carrying out salient region division based on frequency spectrum residual error on the two filtered and equalized images; 4) calculating a minimum euclidean distance between the saliently divided region of the two images by utilizing a histogram mode-sorted grayscale sequence; 5) calculating registration parameters for a matching region by utilizing an improved comprehensive learning particle swarm optimization algorithm; 6) converting the registration parameters into registration parameters of an original image pair by utilizing a coordinate relationship between a sub-region in which a pixel point is positioned and an original image to obtain final registration results. The SAR image registration method solves the registration problem under a situation of non-ideal region division consistence through an existing method and the problem of large registration calculation amount based on grayscale information, and is used for change detection, terrain classification, map correction, data fusion and the like.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, uses similar area information in two images for registration, and is the basis for applications such as change detection, ground feature classification, map correction, and data fusion; SAR image registration method. Background technique [0002] Synthetic Aperture Radar (SAR) has the characteristics of all-weather, all-time, cloud and surface penetration, and is widely used in hydrology, geological mines, environmental monitoring, military and other fields. However, due to the different shooting conditions of the original data, there are relative geometric differences between images, and image registration technology to correct image differences becomes a prerequisite for further image processing. [0003] Based on the idea of ​​​​region-based registration, most of the two images are first segmented, and then the segmented regions of the two images are matched, and then the con...

AI Technical Summary

Problems solved by technology

When the consistency program of the segmented area is not high, it will bring interference to the regional matching of geometric information such as Hu invariant moments, edge chain codes, area area, and perimeter length that describe the area shape information, and based on the area structure or boundary The center of mass and the control points of the polygon fitting will not correspond well, which will affect the registration effect of the above method

For the SAR images taken by the same SAR system at different times, due to the different time points, the same segmentation method with the same parameter settings is very likely to have inconsistent segmentation results.

Images