Abstract:

To provide experimental basis for estimation of total bacterial count in pork using electronic nose, mathematical models describing the relationship between output signal of electronic nose indicating change in volatile composition of chilled pork stored at 4 or 20 ℃ and total bacterial count in pork were investigated by partial least squares regression (PLS-regression). Results showed that pork samples stored for different days could be discriminated by principal component analysis (PCA). The correlation coefficients of developed mathematical models for estimating total bacterial count in pork stored at 4 or 20 ℃ were 0.9003 and 0.9940, respectively, which revealed good linearity. Electronic nose can accurately detect the changes in volatile composition and total bacterial count in pork during storage, thereby providing a useful tool for detecting harmful microorganisms in pork.

Key words: pork, electronic nose, microorganism, detection