Multidimensional discrete filtering using combined discrete Fourier transform and linear difference equation filter methods

Two widely followed approaches to multidimensional (M-D) digital filtering are M-D discrete Fourier transform (DFT) filtering and M-D linear difference equation (LDE) filtering. The authors describe an alternate filtering technique that combines some of the desirable aspects of both methods. The M-D DFT/LDE filtering technique applies DFT filtering to the input signal over less than all of the M dimensions and then applies LDE filtering over the remaining dimensions. Two-dimensional DFT/LDE fan-stop filters, used to remove ground roll from seismic images, are computationally more efficient than their LDE counterparts while providing comparable input-output behavior.<
>