A Fast Planar Homography Estimation Method Based on Multiple Simple Matrix Multiplication

Abstract

The invention relates to a quick planar homography estimation method based on multiplication of multiple simple matrixes. The method comprises the following steps of extracting corresponding feature point pairs between images; selecting two point pairs from the obtained corresponding feature point pairs as reference point pairs, and separately performing two-dimensional similarity transformation and elementary row transformation on points in the images; selecting a third point pair and a fourth point pair from the obtained corresponding feature point pairs except the reference point pairs, and separately solving two plane displacements by using the third point pair; solving two parameters of hyperbolic rotation scaling transformation by using the fourth point pair; changing the fourth point pair, obtaining multiple groups of parameters of hyperbolic rotation scaling transformation by using the changed fourth point pair, and judging the accuracy of the multiple groups of parameters to obtain a planar homography matrix between the images. The method can solve the problems of large calculation quantity, low efficiency and invalidation in the presence of high proportional exterior points.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a fast plane homography estimation method based on multiplication of multiple simple matrices. Background technique [0002] Nowadays, with the development of information technology and the popularization of cameras, computer vision technology is increasingly appearing in people's daily life. Two-dimensional projective transformation, also known as planar homography, is a theoretical basis of computer vision, which describes the mapping relationship between a plane and the imaging surface under the perspective projection of the camera. Planar homography computing is widely used in panorama stitching, 3D reconstruction, vision-based positioning, object detection and other tasks, and is a very common and important technology. [0003] Researchers have proposed various methods to solve the computational problem of planar homography. The current mainstream algorithm uses th...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a fast plane homography estimation method based on multiplication of multiple simple matrices, which can avoid the problems of large amount of calculation, poor efficiency, and failure in the case of high proportion of outliers

Images