Image segmentation of microscopic wastewater images using phase contrast microscopy

Image processing and analysis is a useful tool for monitoring of activated sludge wastewater treatment plants. However, its effectiveness is dependent on performance of the segmentation algorithms. The activated sludge wastewater plant can be monitored by image processing and analysis of images acquired through microscope using bright field microscopy and phase contrast microscopy. In this paper, we have investigated three segmentation algorithms which are channel based segmentation, edge based segmentation and Bradley based segmentation. The performance of the algorithms is assessed using the performance metric of accuracy. Forty gold approximations of ground truth images are manually prepared for comparing with the result for segmentation. Half of the forty images are acquired at 10x and rest at 20x objective magnification of the microscope. Edge based segmentation gives better results compared to other algorithms with accuracy of 0.972.