Regularising 3D Data Using Fourier Reconstruction and Sparse Inversion

A-015 REGULARISING 3D DATA USING FOURIER RECONSTRUCTION AND SPARSE INVERSION P. M. ZWARTJES and C. O. H. HINDRIKS Delft University of Technology Lab. of Acoustic Imaging and Sound Control PO Box 5046 2600 GA Delft The Netherlands Abstract Fourier reconstruction yields a better regularisation than can be obtained by a simple binning procedure. The method consists of estimating the Fourier coefficients by least-squares inversion with a priori information. The sparse distribution of energy over the Fourier coefficients can be used as a priori information. Resulting algorithms perform significantly better in a 3D synthetic experiment than conventional Fourier reconstruction with a