Activated Sludge Image Analysis Data Classification: an LS-SVM Approach

Authors:	Gins Geert, Katholieke Universiteit Leuven, Belgium Smets Ilse, Katholieke Universiteit Leuven, Belgium Jenné Rika, Katholieke Universiteit Leuven, Belgium Van Impe Jan F.M., Katholieke Universiteit Leuven, Belgium
Topic:	8.3 Modelling & Control of Environmental Systems
Session:	Modeling and Control of Wastewater Treatment Plants
Keywords:	classification, complex systems, image analysis, water pollution, waste treatment

Abstract

In this paper, a classifier is proposed and trained to distinguish between bulking and non-bulking situations in an activated sludge wastewater treatment plant, based on available image analysis information and with the goal of predicting and monitoring filamentous bulking. After selecting appropriate activated sludge parameters (filament length, floc fractal dimension and floc roundness), an LS-SVM approach is used to train a classification function. This classification function is shown to have a satisfactory performance after validation. Copyright 2005 IFAC.