A comparative study of two fast phase unwrapping algorithms

作者:

Highlights:

摘要

Let u be a real function defined on a rectangular grid. The (two-dimensional) phase unwrapping problem consists in the reconstruction of u from the knowledge of its values modulus 2π. We study two different methods for the solution of this problem: the Fourier transform method and the network optimization method. We show that these methods do not compute the same solution, so that we introduce a revised version of the Fourier transform method which is equivalent to the network optimization method. These results are based on an explicit expression for a singular value decomposition of A, where A is the matrix that defines the linear constraints of the optimization problem arising in the network optimization method.

论文关键词:Phase unwrapping problem,Minimum cost flow problem,Discrete Fourier transform,Singular value decomposition

论文评审过程:Available online 22 February 2003.

论文官网地址:https://doi.org/10.1016/S0096-3003(02)00920-7