Squeezed light field physical reconstruction method

Abstract

The invention provides a squeezed light field physical reconstruction method. The method comprises steps: (1) multiple random matrixes are generated through a computer, the random matrixes form a large matrix, the large matrix is printed on a film, and a physical mask is generated; (2) the physical mask is used as an acquisition object, a mask image is acquired through moving an aperture arrangedin front of a camera, each mask image is a projection matrix pi, and a measurement matrix phi is acquired through the acquired projection matrix pi; (3) the physical mask is placed before an imaging lens of the camera to form a mask camera, and the mask camera is used to photograph a coded image to be used as a measurement value y; (4) a sparse matrix psi and a sparse vector alpha are acquired; and (5) reconstruction of original signals f is carried out through compressed sensing. The mask is placed before a common industrial camera lens, and through one-time photographing, multi-angle of viewlight field reconstruction can be realized. In comparison with a mode of placing the mask between the camera lens and a CCD, the method of the invention is better in convenience, and the mask camerais realized more easily.

Description

technical field [0001] The invention relates to a reconstruction method of a light field, in particular to a physical reconstruction method of a compressed light field. Background technique [0002] The sum of the radiation functions of each point and each direction in space is the light field. However, the data of the light field is very large, which brings a lot of inconvenience to the storage and transmission of the light field. In 1936, Gershun proposed the concept of light field. In 1992, Adelson proposed the theory of full light field under computer vision. In 1996, Levory parameterized the light field. [0003] It means that the current methods of collecting light field mainly include the microlens array handheld light field camera invented by Ng, the camera array of Stanford University, and the mask light field camera. The camera array adopts multiple cameras to shoot the same target at the same time, but its structure is huge and the photos are expensive. The micr...

AI Technical Summary

Problems solved by technology

These two light field acquisition methods are faced with many inconveniences caused by the storage, transmission and calculation of huge light field data.

Images