Colony counting method based on computer vision research

Computer technology has been widely used in food microbial detection. In this study, we established a method for identifying and counting microbial colonies based on computer vision technology, in order to provide efficient, accurate and reliable data support for microbial detection. The results showed that the deviation of the Baird-Parker plates counting by the computer automatic counting method and the artificial counting method was less than 10%, but the difference of the PCA plates counting results was significantly larger. Paired T test analysis showed that there was no statistically significant difference between the two counting results. As the number of colonies increased, the counting deviation of the two methods gradually increased.