Heterogeneous registration method based on multi-feature constraints

Abstract

The invention belongs to the technical field of image processing, and discloses a heterogeneous registration method based on multi-feature constraints. The method includes the following steps: obtaining a SAR image to be registered and a reference image; performing noise reduction processing on the SAR image to be registered, and reducing Corner detection is performed on the noisy SAR image to be registered and the reference image to obtain multiple first corners and multiple second corners; multiple first corners and multiple second corners are based on the corner position Multi-feature joint constraints of , regional mutual information and distance to obtain multiple pairs of precise matching corner pairs; use RANSAC strategy to iteratively fit multiple pairs of precise matching corner pairs to determine forward control point pairs; use all forward control points For the calculation of precise mapping parameters, the mapping parameters are used to register the SAR image to be registered after noise reduction to the reference image to obtain the registered SAR image. The invention can realize the accurate estimation of the transformation matrix, thereby effectively improving the registration accuracy.

Description

technical field [0001] The present invention relates to the field of image processing technology, in particular to a heterogeneous registration method based on multi-feature constraints, which is used for feature point extraction and feature matching of SAR images and optical images, and can effectively solve the problem of difficult extraction of homonymous points in heterogeneous registration The precise estimation of the transformation matrix can be realized under the following conditions, so as to effectively improve the registration accuracy. Background technique [0002] Heterogeneous registration is to map heterogeneous data with spatiotemporal misalignment and inconsistent granularity into spatiotemporal consistent data. These spatiotemporal consistent data are the basis for subsequent multi-source association detection, joint identification and information fusion. Because SAR images usually have low signal-to-noise ratio and defocusing, it is extremely difficult to ...

AI Technical Summary

Problems solved by technology

However, it is difficult for the existing method to accurately match the features in the process of heterogeneous registration, and it is difficult to estimate the fine transformation matrix, which makes the registration accuracy of the existing method low

Images