RESUMO
The measurement of microbial contamination is of primary importance in different fields, from environmental monitoring to food safety and clinical analysis. Today, almost all microbiology laboratories make microbial concentration measurements using the standard Plate Count Technique (PCT), a manual method that must be performed by trained personnel. Since manual PCT analysis can result in eye fatigue and errors, in particular when hundreds of samples are processed every day, automatic colony counters have been built and are commercially available. While quick and reliable, these instruments are generally expensive, thus, portable colony counters based on smartphones have been developed and are of low cost but also not accurate as the commercial benchtop instruments. In this paper, a novel computer vision sensor system is presented that can measure the microbial concentration of a sample under test and also estimate the microbial growth kinetics by monitoring the colonies grown on a Petri dish at regular time intervals. The proposed method has been in-house validated by performing PCT analysis in parallel under the same conditions and using these results as a reference. All the measurements have been carried out in a laboratory using benchtop instruments, however, such a system can also be realized as an embedded sensor system to be deployed for microbial analysis outside a laboratory environment.