Water quality analysis: A pattern recognition approach

Abstract The development of a microcomputer-based system for the automatic counting of colonies of indicator organisms (that is, bacteria whose presence is indicative of the presence of pathogens) trapped on membrane filters from water samples is described here. Three different approaches, based on mathematical morphology, distance transform and fuzzy c-means clustering respectively, are implemented and their relative performances are discussed here.