Image generation method based on internal-parameter-corrected fisheye cameras

Abstract

The invention relates to an image generation method based on fisheye cameras. An original plane model for imaging is provided. The method comprises a view spherical model building step in which view spherical models corresponding to fisheye cameras are built according to initial external parameters; an internal parameter correcting step in which the benchmark mapping relationship between the view spherical model of each fisheye camera and a benchmark spherical model is determined according to initial external parameters; an external parameter correcting step in which the same second feature points in the benchmark view spherical model are determined as a same-name group, the deviation distance between the second feature points in each same-name group is calculated, the initial external parameters of each fisheye camera are corrected until the sum of deviation distances of all the same-name groups is minimum, and thus corrected external parameters are obtained; and a panoramic imaging step in which images collected by the fisheye cameras are projected to the benchmark spherical model according to the corrected internal parameters and the corrected external parameters to get a panoramic image.

Description

technical field [0001] The invention relates to an image processing method, in particular to an image generation method for a fisheye camera based on internal parameter correction. Background technique [0002] Panoramic video is a low-cost virtual reality technology based on image stitching technology, and it is a research hotspot in the field of virtual reality and computer vision technology. Panoramic stitching technology has been widely used in many fields as a cheap and intuitive way, such as VR / AR, video shooting, video surveillance, real estate, decoration design, tourist attractions, virtual campus, street view services, etc. market demand. In order to capture larger and more complete high-definition panoramas, a panoramic camera usually consists of more than two cameras to cover a full range of viewing angles. Since the lens of a fisheye camera has the characteristics of a large field of view, the panoramic camera Most of them are composed of fisheye cameras. [...

AI Technical Summary

Problems solved by technology

[0003] To generate a panoramic image, it is necessary to splicing the collected original images through several fisheye cameras. If splicing is to be realized, the precise internal parameters and external parameters of each fisheye camera must be obtained; However, due to the limitations of industrial manufacturing technology, the relative pose relationship between cameras cannot meet the accurate alignment requirements of the theoretical value of the pose relationship between cameras. Stitching of acquired video images will produce obvious misalignment and ghosting

Images