A Pistachio Nuts Classification Technique: An ANN Based Signal Processing Scheme

This paper introduces an intelligent system for pistachio nuts classification using artificial neural networks (ANNs). The employed ANN is trained based on analyzing of acoustic signals, generated from pistachios impacts with a steel plate. The fast Fourier transform (FFT), discrete cosine transform (DCT), and discrete wavelet transform (DWT) are employed and compared for processing the signals. In order to restrict the signals dimensions, principle component analysis (PCA) algorithm is used. The experimental results are demonstrated for various types of ANNs with different number of hidden layers and neurons to establish optimum result. The results demonstrate feasibility and performance of the proposed method with an accuracy of more than 99.89%.