The problem of detecting leakages in pipelines has been assuming a crucial importance for saving water from bad usage and exploitation. FDM (Filter Diagonalization Method) is a promising method used in nuclear magnetic resonance for decaying issues. We propose a comparison between a one-dimensional FDM and a multidimensional one. This comparison points out interesting aspects related to the leak width and noise contained in the signal and that are hidden in 1 D FDM. An interesting result is the fact that, thanks to multidimensional approach, finding other peaks and lobes, it is possible to improve the construction of a pipelines and to reduce the number of sensors to be located on the pipeline.