Development of an image analysis procedure for identifying protozoa and metazoa typical of activated sludge system

A procedure for the semi-automatic identification of the main protozoa and metazoa species present in the activated sludge of wastewater treatment plants was developed. This procedure was based on both image processing and multivariable statistical methodologies, leading to the use of the image analysis morphological descriptors by discriminant analysis and neural network techniques. The image analysis program written in Matlab has proved to be adequate in terms of protozoa and metazoa recognition, as well as for the operating conditions assessment.