An image registration method based on double feature detection

Abstract

The invention discloses a twice feature detection based image registration method which comprises the following steps that when feature detection is performed on an image, a certain quantity of feature points are extracted firstly, a mask image is generated with feature point coordinates and neighborhoods thereof and is utilized to perform secondary feature detection, the feature points obtained by the two detections are combined as a final feature detection result to be output, and then operations including feature description, feature matching, transformation model estimation and image interpolation to obtain a registrated image. According to the twice feature detection based image registration method provided by the invention, by utilizing the position relation among the feature points,a certain quantity of feature points are still kept in a region with weaker feature strength, richer feature point information is provided, a correctly matched amount in a feature matching link is increased, and high-accuracy image registration can be realized under the conditions of poorer imaging condition, target internal structure change and the like; and meanwhile, the twice feature detection based image registration method can be adapted to various feature detection and feature description operators, omits a complex parameter regulation process and has conciseness.

Description

technical field [0001] The invention belongs to the field of image processing, and more specifically relates to an image registration method based on two feature detections. Background technique [0002] Image registration is the process of superimposing two or more images of the same scene at different times, different viewing angles, and different sensors. It realizes the geometric alignment of the reference image and the image to be registered. Image registration has a wide range of applications, including machine vision, 3D reconstruction, remote sensing image processing, object classification and retrieval, image understanding and fusion, etc. [0003] Image registration can be divided into three categories: region-based, frequency-domain-based and feature-based. The features in the image mainly include feature points, edges, and contours, and the feature points are often points in the image that have large grayscale changes in all directions, including inflection poin...

AI Technical Summary

Problems solved by technology

[0006] In view of the above defects or improvement needs of the prior art, the present invention provides an image registration method based on two feature detections, thereby solving the problem of poor imaging conditions, less prominent feature areas, and changes in the internal structure of the target in the prior art. There is a technical problem of low image registration accuracy in the image

Images