Akademik Veri Yönetim Sistemi
Elektronika ir Elektrotechnika, cilt.30, sa.5, ss.4-13, 2024 (SCI-Expanded)
An acoustic emission and machine learning-based system for pistachio classification has been developed, utilising Mel frequency cepstral coefficients (MFCC) for feature extraction and a support vector machine (SVM) for classification. The study found that closed-shell pistachios produce different frequency components than open-shell pistachios when they hit a steel plate. Audio signals were recorded using a high-sensitivity carbon microphone and a MATLAB Analogue Input Recorder. These recordings were processed with a Hamming window to reduce ambient noise. MFCCs, a leading method for extracting audio signal features, were used to differentiate between open- and closed-shell pistachios. The extracted features were input into the fit classifier support vector machine (FITCSVM) algorithm for classification, which performs binary classification on low- or medium-sized data sets. The study achieved high accuracy in distinguishing between open- and closed-shell pistachios, highlighting the potential of this system for the pistachio industry to improve product quality and processing efficiency. In conclusion, the MFCC and support vector machine (SVM) algorithms effectively classified the pistachios by analysing acoustic emissions. This innovative approach shows promise in the development of more efficient methods in the processing of agricultural products.