Image matching method based on improved GMS-ORB feature and storage medium

Abstract

The invention requests to protect an image matching method based on improved GMS-ORB features, and the method comprises the following steps: S1, carrying out the rough extraction of the feature pointsof an original image and a to-be-registered image through employing a Fast method, and obtaining a large number of feature points; S2, using a machine learning method to screen out optimal Fast feature points; S3, constructing a Gaussian pyramid to realize scale invariance of the feature points; S4, realizing rotation invariance of the feature points by using a gray centroid method; S5, describing the feature points extracted from the Fast corner points by using a BRIEF algorithm, and performing feature point rough matching by using a violent matching algorithm based on Hamming distance; andS6, purifying the rough matching by using an improved GMS algorithm, and removing wrong matching to complete image registration. According to the method, the quadtree and the GMS algorithm are combined, the quadtree is used for controlling the number of the grids generated by the GMS, the matching time is shortened, the grids are divided in a targeted mode, and the registration precision is improved.

Description

technical field

[0001] The invention belongs to the technical field of image processing, in particular to an ORB feature extraction method, a GMS feature point purification method and a combination of three methods of quadtree segmentation. Background technique

[0002] Images are an important source of people's cognition of themselves and the world. With the rise of computer vision, image matching technology has attracted more and more attention as the basis of image processing. At present, many famous scholars at home and abroad are committed to the research of image matching algorithms. There are three main types of image matching methods. The first type is image matching algorithms based on grayscale areas. This type of algorithm mainly operates on the grayscale of image pixels. Calculate the correlation between the geometric sub-regions of the image to perform matching, so as to complete the matching; the second type is an image matching algorithm based on the transform...

Images