Melamine is a chemical compound generating a very characteristic Raman signature. This component is fraudulently added to milk to artificially increase its nitrogen content (and thereby its apparent protein content). In this paper, we evaluate the ability of a specific wide-field Fourier-based Raman imaging system to identify melamine. The melamine is studied in powder form or diluted and dried out on SERS (surface-enhanced Raman spectroscopy) Cu2O-Ag substrate, which enhances the Raman signature. In both forms, we show that the spatial content of the information is an asset to characterize the samples; on SERS, we demonstrate the inhomogeneity of the deposition, and in powder form, we identify melamine as individual grain mixed with milk powder. The specific wide-field imaging technology features much higher laser excitation with lower local intensity than traditional point to point Raman measurement, thereby reducing the acquisition time for full data sets.