Onaran, İbrahimDülek, BerkanPearson, T. C.Yardımcı, Y.Çetin, A. Enis2016-02-082016-02-082005http://hdl.handle.net/11693/27334Date of Conference: 4-8 September 2005Conference Name: 13th European Signal Processing Conference, EUSIPCO 2005Shell-kernel weight ratio is the main determinate of quality and price of hazelnuts. Empty hazelnuts and nuts containing undeveloped kernels may also contain mycotoxin producing molds, which can cause cancer. A prototype system was set up to detect empty hazelnuts by dropping them onto a steel plate and processing the acoustic signal generated when kernels impact the plate. The acoustic signal was processed by five different methods: 1) modeling of the signal in the time domain, 2) computing time domain signal variances in short time windows, 3) analysis of the frequency spectra magnitudes, 4) maximum amplitude values in short time windows, and 5) line spectral frequencies (LSFs). Support Vector Machines (SVMs) were used to select a subset of features and perform classification. 98% of fully developed kernels and 97% of empty kernels were correctly classified.EnglishAcoustic signalsComputing timeFrequency spectraImpact acousticsLine spectral frequenciesMaximum amplitudePrototype systemSteel platesTime domainTime windowsAcoustic wavesSignal detectionSupport vector machinesConference Paper