Surf algorithm based quick mismatching point eliminating method

Abstract

The invention relates to a technology for eliminating mismatching points between two images and utilizing correct matching point pairs to compute an image transformation matrix, in particular to a surf algorithm based quick mismatching point eliminating method, in order to avoid the defects of large computation amount and high mismatching ration in the matching process and realize quick and efficient mismatching point pair elimination and image transformation matrix computation. The method includes: for a surf algorithm, computing each feature point including information of coordinates (x, y, S) of the feature point, wherein S represents scale space of the feature point; for two to-be-matched points extracted from the two images, screening according to the feature points of the matching points, and respectively setting a scale limit threshold value and a main direction limit threshold value for screening; performing corresponding computation and judgment for a main direction angle; after matching point pairs of a required number are found, performing a fine matching process. The method is mainly applied to elimination of the mismatching points between the two images.

Description

technical field [0001] The invention relates to a technique for efficiently and quickly eliminating wrong matching point pairs between two images and calculating image transformation matrix by using correct matching point pairs. Specifically, it relates to a method for quickly eliminating mismatching points based on the surf (Speeded Up Robust Features) algorithm. technical background [0002] Today, the research of image stitching technology is an increasingly popular research field, and it has a wide range of application values ​​in photogrammetry, computer vision, remote sensing image processing, medical image analysis, computer graphics and other fields. Generally speaking, the process of image stitching consists of three steps: image acquisition, image registration, and image fusion, and image registration is the basis of the entire image stitching. Image stitching has been one of the key points in image research in recent years, and researchers at home and abroad have...

AI Technical Summary

Problems solved by technology

In general registration algorithms, the matching process of feature point pairs is divided into two processes: rough matching and fine matching. A pair of feature points with the smallest distance is regarded as a matching point pair. In fine matching, the method of randomly extracting matching point pairs is used to calculate the transformation matrix and the number of internal points, which will cause the following problems in the matching process. In the rough matching process The minimum Euclidean distance obtained in , and the nearest neighbor point is not necessarily the correct matching point when the second nearest neighbor point is also small, which greatly increases the false matching rate, and for each matching point to be calculated once 64 dimensions The Euclidean distance of the vector is undoubtedly a large calculation overhead, and in the fine matching, it is necessary to ensure that a more accurate transformation matrix is ​​obtained, and the way of randomly picking points also greatly increases the number of point picking, which also increases the matching process. amount of computation

Images