Abstract

 Monitoring food safety is so important because the tainted food could do harm to people. The main methodology to check the food safety is to measure the number of microbes which is the cause of decomposition. There are several method to determine the number of microbes. But existing ones are too slow or too inaccurate. Proposed method provides the fast and more accurate method based on the image analysis. To determine the number of microbes in food sample, the digital images of the microbes which are separated from the food through a series of process and magnified by microscope are taken. We take the images as the input of analysis system and measure the number of microbes in the images. And a statistical analysis on the results gives the more accurate density of microbes in the food.

 To count microbes is to detect and classifying microbe in the image. By some image processing, we can get the colony regions of microbes in the image. The regions are classified into three classes of microbe - bacillus, coccus, yeast.

 Efficiency of the proposed method was shown by the experiment. The proportion of the estimated number to the real number from the experiment was in the range of 0.80 and 1.66, while the applicable system to estimate the number of microbe should have the proportion in the range of 0.1 and 10.0.

Related Knowledge

 There are several type of microbes. Figure 1 shows the types of microbes which are classified by shapes.

Notes

Participants: Younghee Kwon, Sungho Ryu
Project ended in 2001