Rapid Measurement Of Sesame Quality By Means Of Near-infrared Spectscopy Technique

Rapid non-destructive measurement of sesame quality is an important technical that toensure pricing sesame according to quality during acquisition, storage. In order to researchrapid measurement of sesame quality by near infrared spectroscopy, the subject collected66kinds of white sesame,62kinds of black sesame seeds from across the country (mainly in theCentral Plains), and selected sesame quality indicators such as moisture, ash, protein, crudefat, and sesamin to measured and analyzed.Collecting sesame whole spectral acquisition inthe range of570nm～1100nm using near infrared grain analysis instrument. The factors thataffecting the model results: light Scattering processing, mathematical processing parametersand regression techniques were Inspected, with chemical determination of the value of qualityindicators. Then we selected the best conditions of the model parameters and established thequantitative detection of sesame quality model.The near-infrared prediction models of white sesame moisture, ash, protein, crude fatand sesamin were established by using Modified PLS, with None, Inverse the MSC, None,Detrend only and Standard MSC Were selected,"2,4,4,1","3,10,10,1","1,4,4,1","4,3,3,1"and "3,4,4,1" mathematical treatment were applied respectively. The RSQ were0.9964,0.9124,0.9786,0.9949and0.8808respectively. According to the internal cross-validationand external validation,1-VR were0.9381,0.9095,0.9447,0.9759and0.8788, R~2were0.942,0.920,0.968,0.973and0.834respectively. The RSD of five indicators were all lessthan10%. The RPD of moisture protein and crude fat were all greater than3, the RPD of ashwas greater than2.5and less than3, and the RPD of sesamin was less than2.5.The near-infrared prediction models of black sesame moisture, ash, protein, crude fatand sesamin were established by using Modified PLS (except crude fat, PLS), with Detrendonly, Weight MSC, Detrend only, SNV+Detrend and Detrend only Were selected,"2,3,3,1","3,10,10,1","1,3,3,1","2,3,3,1" and "3,1,1,1" mathematical treatment were appliedrespectively. The RSQ were0.9304,0.8935,0.9166,0.9703and0.8758respectively.According to the internal cross-validation and external validation,1-VR were0.9068,0.8861,0.9043,0.9221and0.8306, R~2were0.902,0.879,0.913,0.917and0.714respectively. TheRSD of four indicators were all less than10%(except sesamin). The RPD of moisture proteinand crude fat were all greater than3, the RPD of ash was greater than2.5and less than3, andthe RPD of sesamin was less than2.5. The model of ash was used to roughly measure because of poor accuracy. The model ofsesamin was not used to measure because RPD was less than2.5. The model of moisture,protein and crude fat were used to rapid mesurement because of great calibration results.These models could determine the moisture, protein and crude fat content in sesame used as arapid method to detect the quality of sesame seeds.