Large change density fringe pattern filtering method based on generalized total variation

Abstract

The invention belongs to the optical detection and optical information processing technology field, and relates to a large change density fringe pattern filtering method based on generalized total variation. A new variation image decomposition model is provided, and is used in optical fringe pattern processing, and then a thick fringe part and a thin fringe part are separately processed, and in addition, the smoothness of the thick fringe part is guaranteed, and at the same time, the thin fringe part is not dimmed, and therefore a good large change density fringe filtering effect is provided. By adopting the technical scheme provided by the invention, the large change density fringe pattern filtering method based on the generalized total variation is characterized in that a second-order generalized total variation is used to describe the low density part, and Hilbert space is used to describe the high density fringe, and simple L2 space is used to describe a noise part, and then by optimizing a functional formula, every component u,v,w is finally acquired. The large change density fringe pattern filtering method is mainly used for optical detection and optical information processing occasions.

Description

technical field [0001] The invention belongs to the technical field of optical detection and optical information processing, and relates to a filtering method based on generalized total variation for filtering large-variation density fringe images. Background technique [0002] At present, great progress has been made in the filtering method of electronic speckle interference fringe pattern, but there are still some difficulties in processing the fringe pattern with large variation density. Bidirectional PDE filtering method and window Fourier method are two effective fringe image filtering methods, but it is difficult for them to ensure that thick fringes are smooth while fine fringes are not blurred when dealing with fringe images with large variation density [1, 2]. In recent years, variational image decomposition algorithms have received extensive attention from researchers and have developed into a promising and very novel research field. Variational image decompositio...

AI Technical Summary

Problems solved by technology

For example, the proposal of the adaptive full score model [4] is based on a threshold parameter given in advance, and this threshold parameter is difficult to choose

Although the results obtained by the non-convex function variational model [5] are better, but the global optimal solution cannot be obtained

Images