| The odor information of vacuum -packed tilapia during refrigerated storage were captured by electronic noses, its recognition sensors and the feature extraction were analyzed based on fish quality change, to aim at screening the suitable sensor types and feature extraction for distinguishing fish quality. The results showed that the response signal of some sensors increased with the enhancement of the odor concentration during the storage, which fish freshness was decreased gradually. The signal of 40~42 s was stable,and was more consistent with quality change trends, it could be used for feature extraction; the sensors types with good linear relationships of quality indictors were obtained by using correlation analysis and stepwise regression analysis; these screened sensors and feature extraction were used for pattern recognition, it could well distinguish different fish quality grades, indicating that the screening conditions of sensors could be used for quality determination. |
