Effective object recognition for automated counting of colonies in Petri dishes (automated colony counting).

Determination of the number of colonies (colony forming units, CFU) is a standard method in microbiological analysis to ensure the quality of drinking water. Normally this tedious work is still performed manually. A PC-based method for the automated counting of digitized images of Petri dishes is presented. The method includes highly specific and effective object recognition algorithms that ensure very high detection accuracy. The processing sequence implies internal controls therefore enabling reliable automated evaluations of series of images. Use of the Fuzzy formalism and the high adaptivity of the algorithms lead to an extraordinary user-friendliness. For digitization different devices like flatbed scanners or CCD-cameras can be used. Due to the highly adaptive algorithms samples from the routine standard preparation process in laboratories can be evaluated. The accuracy and quality of the method aim at advancement in objectivity of colony counting and quality control and assurance. The algorithms and the evaluation of the method are presented.