Automatic determination of bacterioplankton biomass by image analysis

Image analysis was applied to epifluorescense microscopy of acridine orange-stained plankton samples. A program was developed for discrimination and binary segmentation of digitized video images, taken by an ultrasensitive video camera mounted on the microscope. Cell volumes were estimated from area and perimeter of the objects in the binary image. The program was tested on fluorescent latex beads of known diameters. Biovolumes measured by image analysis were compared with directly determined carbon biomasses in batch cultures of estuarine and freshwater bacterioplankton. This calibration revealed an empirical conversion factor from biovolume to biomass of 0.35 pg of C mum (+/- 0.03 95% confidence limit). The deviation of this value from the normally used conversion factors of 0.086 to 0.121 pg of C mum is discussed. The described system was capable of measuring 250 cells within 10 min, providing estimates of cell number, mean cell volume, and biovolume with a precision of 5%.