Method for three-dimensional reconstruction of multilevel lens multi-view scene

Abstract

The invention discloses a method for three-dimensional reconstruction of a multilevel lens multi-view scene. The method comprises the following steps of: acquiring a multi-view two-dimensional image by a multilevel lens three-dimensional imaging system; performing three-dimensional reconstruction on the multi-view two-dimensional image by using a holographic reconstruction method, a stereoscopic vision method or a method with the combination of stereoscopic vision and holographic reconstruction to obtain depth information of a real image of an object in the scene through the last level lens; and finally, performing progressive recursion on the depth information of the real image along the direction which is opposite to that of a light ray entering the multilevel lens three-dimensional imaging system according to a relational expression of the focal distance, the object distance and the image distance of a lens, thereby obtaining three-dimensional information of the object in the scene.

Description

technical field [0001] The invention belongs to the technical field of two-dimensional or three-dimensional imaging, and more specifically, the invention relates to a method for three-dimensional reconstruction of a multi-level lens multi-view scene. Background technique [0002] Computer vision is a very active field in computer science. 80% of the sensory information obtained by humans comes from vision, and computer vision is to use cameras and computers to simulate human stereo vision to obtain targets and analyze them. Stereo vision matching technology is the most critical research branch in the field of computer vision, and it is a passive three-dimensional imaging technology based on the principle of triangulation. Passive 3D imaging technology does not require the introduction of external light sources, but for images without obvious features (such as no obvious edges, corners, or no texture and repeated textures, etc.), the amount of calculation is very large, and ...

AI Technical Summary

Problems solved by technology

[0014] In order to solve the problem that the matching result of stereo vision is difficult to predict for scenes without significant texture or texture repetition, the purpose of the present invention is to use the feature of holographic reconstruction that does not have high requirements for scenes, and use the method of holographic reconstruction to carry out the reconstruction results of stereo vision. Overall correction or local correction, which solves the problem that stereo vision is difficult to match with scenes without significant texture or texture repetition

Images