Panoramagram generation method and system based on characteristic matching

Abstract

The invention provides a panorama generation method based on character matching and a system. The method comprises the following steps: characteristic points of an image A and an image B which are adjacent are extracted, vector parameters of the characteristic points are established, and coordinates of the characteristic points are determined; the characteristic points of the image A are used as a sample, the characteristic points of the characteristic points are classified, and the characteristic points which are the best matching with the sample in the image A are selected out from each classification of the image B; a relational expression for the image A to be converted to be under the coordinate system of the image B is calculated by utilizing the matching relations of the characteristic points, and the image A is converted to be under the coordinate system of the image B according to the relational expression; and the image A and the image B are spliced to form an image A1. The panorama generation method and the system use a nearest neighbor method training classifier to match the characteristic points, and obtain a conversion relational-expression with high accuracy through a total least square method, thereby generating a panorama rapidly and accurately.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a method and system for generating panoramas based on feature matching. Background technique [0002] The viewing angle range is an important parameter of a camera device, and is used to measure the range of scenes that the camera device can capture. The viewing angle range of general camera equipment is very limited, about 50 degrees in the horizontal direction and about 35 degrees in the vertical direction; the viewing angle range of the human eye is relatively wide, about 200 degrees in the horizontal direction and about 135 degrees in the vertical direction; however, the viewing angle range in the entire 3D scene , which is 360 degrees in the horizontal direction and 180 degrees in the vertical direction. Obviously, the information of the entire 3D scene cannot be obtained at the same time, no matter through the camera equipment or the naked eye. To solve th...

AI Technical Summary

Problems solved by technology

However, in the panorama generation method based on feature matching, when applying the traversal search method to perform feature matching on feature points, the amount of calculation is very large, and some matching feature points are easy to miss, which affects the efficiency and accuracy of panorama generation. Spend

Images