a novel acoustic sensor approach to classify seeds based on sound absorption spectra一种新型声学传感器基于吸声光谱方法对种子进行分类.pdfVIP

Sensors 2010, 10, 10027-10039; doi:10.3390/s101110027 OPEN ACCESS sensors ISSN 1424-8220 /journal/sensors Article A Novel Acoustic Sensor Approach to Classify Seeds Based on Sound Absorption Spectra Vicent Gasso-Tortajada, Alastair J. Ward, Hasib Mansur, Torben Brøchner, Claus G. Sørensen and Ole Green * Department of Biosystems Engineering, Faculty of Agricultural Sciences, Aarhus University, Blichers Allé 20, P. O. Box 50, DK-8830 Tjele, Denmark; E-Mails: vicent.gassotortajada@agrsci.dk (V.G.T.); alastair.ward@agrsci.dk (A.J.W.); hasib.mansur@agrsci.dk (H.M.); torben.brochner@agrsci.dk (T.B.); claus.soerensen@agrsci.dk (C.G.S.) * Author to whom correspondence should be addressed; E-Mail: ole.green@agrsci.dk; Tel.: +45-8-999-1900; Fax: +45-8-999-1619. Received: 14 September 2010; in revised form: 1 November 2010 / Accepted: 5 November 2010 / Published: 9 November 2010 Abstract: A non-destructive and novel in situ acoustic sensor approach based on the sound absorption spectra was developed for identifying and classifying different seed types. The absorption coefficient spectra were determined by using the impedance tube measurement method. Subsequently, a multivariate statistical analysis, i.e., principal component analysis (PCA), was performed as a way to generate a classification of the seeds based on the soft independent modelling of class analogy (SIMCA) method. The results show that the sound absorption coefficient spectra of different seed types present characteristic patterns which are highly dependent on seed siz