Spectroscopy And Spectral Imaging Technology And Their Applications On Food Detection

Spectra contain abundant physical and chemical information of substances.The spectrum detection technology,combining spectroscopy and chemometrics,is a fast,non-destructive and efficient detection method,which has been widely used in biological,environmental and food detection fields.In recent years,as the food safety issue has been increasingly concerned,it has become a research hotspot to develop variety of spectral analysis techniques and applied them into the detection of food quality.It is quite significant to study in the crossing field among spectroscopy,chemometrics and food science.The main work of this thesis is to study spectral and spectral imaging techniques and corresponding detection systems as well as their applications in food detections.Firstly,the near infrared spectroscopic technique,mostly widely used in the food research field,is investigated.By using principal component analysis,partial least square,and interval partial least square methods,the two types of Acacia gums are qualitatively discriminated and quantitatively analyzed.Moreover,the effects of different spectral pretreatment algorithms are compared and the influence of sampling methods on the test results is discussed for the first time.Secondly,the fluorescence spectroscopic technique,also widely used in food detection with the advantage of low cost and easy integration is studied.A compact and automatic fluorescence spectra detection system based on LEDs is developed for food quality measurements,e.g.,apple and tea detections.For apples'quality attributes detection,the similar prediction accuracy as the near infrared spectroscopy technique is obtained.The accurate prediction of tea grade is also reliable.In order to further increase the detection accuracy,the two-dimensional correlation spectroscopy technique is investigated.A fast two-dimensional fluorescence correlation spectroscopy detection system is proposed and developed for the first time,base on which,a fast,automatic and high accurate detection of Longjing tea grades could be realized.Finally,taking into account of the need for simultaneous detection of sample space and material information in the food industry,a novel point-scan hyperspectral imaging system is developed.Compared to the traditional point-scan system,it is more compact,more accurate and more stable.Using this hyperspectral fluorescence imaging system,the mixture of coffee and coffee mate powder is quantitatively analyzed.All these developed spectroscopic and hyperspectral imaging techniques could be further applied to the study of a wide range of food testing and other related fields.