Acoustic sensor for identifying the frying end-point of chicken steaks

Abstract

The identification of frying end-point of chicken steaks is a prerequisite for developing intelligent equipment and improving product's quality. To achieve this objective, acoustic sensing technology was firstly proposed to classify undercooked chicken steaks (UCSs) and cooked chicken steaks (CCSs). Ensemble empirical modal decomposition (EEMD) was used to reconstruct a new signal by reducing redundant information. Forty-two acoustic features were extracted from time-domain, frequency-domain and time-frequency-domain signals to establish the classification models between the UCSs and CCSs. An optimized classification model with accuracy of 98.84 % was built by LIB support vector machine based on 42 features from EEMD processed signals. The microstructures of UCSs and CCSs, whatever cross-sections or outside surfaces, were significantly different, which probably played an important role in building an excellent classification model. The proposed method consists of non-invasive and reliable technique for online identifying the frying end-point of steaks.