Image analysis as a tool to recognize anaerobic granulation time

Abstract

Image analysis techniques were used to monitor the development of anaerobic granules into an EGSB reactor fed with a synthetic substrate based on glucose (30-40% COD) and volatile fatty acids (70-60% COD) during 400 days. Two types of objects were recognized and quantified: aggregates and free filaments. The aggregates were classified into five ranges based on the equivalent diameter: smaller than 11.3 mm (residual), between 11.3 and 35.7 mm, between 35.7 and 112.8 mm, between 112.8 and 356.8 mm and larger than 356.8 mm. For each size range the following parameters were measured: projected area, number of aggregates, width, length, perimeter, and convex envelop. For filaments, the measurements were: number of free filaments and total free filament length. Other parameters were calculated based upon this data: average equivalent diameter, percentage in number of aggregates, percentage in total area, fractal dimensions and a shape factor (P2A), based on the perimeter and on the projected area. Besides, a new parameter was defined in terms of the ratio of filament length to total floc area (LfA). It showed to be very sensitive to the changes in the morphology of the developing granules, and, based upon our granule definition, it was used to determine the granulation time (θg) in our system (120 days). The P_2A factor allowed the quantification of aggregates roughness and was much more sensitive than fractal dimensions. The study was complemented with measurements of specific methanogenic activity (SMA) of the granules in presence of acetate, propionate, butyrate, ethanol and a gas mixture of hydrogen/carbon dioxide.