This work on imaging of particulate processes using electrical charge tomography uses two basic techniques: the multi-sensing of electrical charge in a cross-section of the flow pipe and a neural network based flow regime identification system to aid in the image reconstruction process. A measurement system, consisting of sixteen sensors, placed at equal distance from each other around the boundary of a circular 100 mm bore pipe, is used to determine the voltage profile of the flow for several artificially produced flow regimes: full, annular, core, half and stratified. A sand flow system is used to produce these different flow regimes, which are created artificially by using baffles of different shapes that obstruct the sand flow. The voltage profile from the sixteen sensors gives spatial information of the flow regime. These profiles are normalised and formed into patterns that are presented to a Kohonen neural network for classification. Two regime classification between clearly distinct regimes gives an accuracy of identification of 85%. Classification of closely similar patterns show much less accuracy of 30%.