Development And Application Of Near-Infrared Spectroscopy Analysis Software Based In C++ And MTALAB For Edible Oil

In recent years, Edible Oil security problem has been occurring. There are many factors leading to this security problem, one important factor is lack of an effective detection and identification techniques of edibleoil authenticity. The traditional method of Gas chromatography and Liquid chromatography can solve the problem of authenticity of Edible Oil in a certain extent, but it has the disadvantages, such as sample handing cumber some and expensibe. Near-Infrared(NIR) Spectroscopy is one of the most commonly used nondestructive testing technologies in recent years. Comparing to the traditional chemical analysis method, it is pollutionfree, easy of ordering, low cost and fast. Now, it is widely used in qualitative & quantitative detection of food and drug. So, this paper developed a Edible Oil quality rapid anslysis software which based on Near Infrared Spectrosopy and chemomrtrics methods and using platform of MATLAB and VC++. The article has carried out research work as follows :(1) This paper introduces the analysis technique of Near Infrared Spectroscopy datas by chemometrics method : pretreatment--- modeling--- forecast. We discussed three kinds of pre-treatment methods and two kinds of qualitative discrimination methods and one quantitative analysis method. All of the methods have been verified by Matlab.(2) Develped the Near-Infrared spectroscopy analysis software based on VC++ platform. The main function of the software include : public module, pretreatment module, calibration model module, model derived module and prediction module.(3) We tested the performance and evaluation of this software,that including qualitative and quantitative testing. We get a good predition accuracy whic h used the qualitative model by Support Vector Machine(SVM) and quantitative model by Support Vector Regression(SVR), both of the models were pre-processing by Standard Normal Variate(SNV).