One of the most noticeable consequences of global climate change is the increased occurrence of algae and cyanobacteria blooms in surface waters. Some of these organisms may release hazardous toxins which represent a threat for human and animal health worldwide. Accordingly, the identification of threshold levels of toxic cyanobacteria cells has become common practice to ensure successful water management. The majority of current methods for cyanobacteria enumeration and bio-volume assessment are very time-consuming and costly. Furthermore, when dealing with multicellular organisms (i.e., filaments, colonies, agglomerates etc.), none of the existing enumeration methods can achieve good accuracy and all tend to underestimate cell concentrations and bio-volume. In this study, we introduce an integrated method for automated enumeration and bio-volume estimation of Anabaena flos-aquae, a common filamentous species of cyanobacteria often present in water blooms. Since Anabaena filaments are often long and tangled, a sample of its culture was first sonicated to isolate individual cells, and then imaged while being excited by a 546nm light source to considerably improve contrast. A probabilistic unsupervised classification was introduced to detect the target cells, and the size distribution of the cells was used for model calibration. Using this learned cell model, subsequent samples with natural Anabaena filaments were automatically enumerated and the bio-volume estimated. Compared to traditional manual enumeration using a hemacytometer, the developed method achieved equivalent accuracy in much less time, with less resources, and provided additional bio-volume information. These preliminary results demonstrate the potential of the developed method as a robust tool for water quality monitoring.