Direct enumeration and size classification of soil bacteria using confocal laser scanning microscopy and image analysis

Direct measurement of fluorescently stained soil bacteria is a common method in microbiology. The tedium of making such measurements visually has prompted the use of automatic image analysis. Accurate measurements by image analysis require an reliable method of image processing that eliminates background particles and selects an appropriate threshold for automated counts and measurements. The authors tested two processing methods, a sharpening filter and a background correction, with subsequent threshold setting and direct threshold setting without previous image processing. The threshold was chosen according to the first maximum of the first derivative of the characteristic line of one sample. The background correction method was found to be superior for analyzing images of soil bacteria, although it leads to slight underestimation of the count as well as the area measurement.