Improving Food Quality Analysis Using a Wavelet Method for Feature Extraction

Measurements of food products can be helpful in the food industry for fast and immediate quality control of the products. An electronic tongue can be used for these measurements since its response contains information about the sample´s total chemical and bacteriological content. The main issue is the data analysis of these measurements. We propose a method based on the wavelet transform together with different selection criteria to choose the most suitable wavelet coefficients for representing the original signal. The proposed method is tested on measurement data for baby food where the purpose was to monitor the ageing process. Using the proposed method provided better overall identification of different stages in the ageing process