Development Of A Portable Near-infrared Spectroscopy Instrument For The Detection Of Adulteration Of Camellia Oil

Food quality and safety is closely related to people’s health and quality of life, so the detection of food quality and safety become an important duty of food regulatory authorities in Global including China. Edible vegetable oil is the necessities of consumer food, and with the increasing demand for quality of life, there is a growing favor for special and functional edible vegetable oil. The camellia oil, which has unique therapeutic effect, is known as the king of oils, and because of its limited production, its price is much higher than other common edible vegetable oils. Based on these above, some unscrupulous traders mix camellia oil with some cheap vegetable oil to obtain illegal gains, so it has great practical significance to develop a set of instrument to detect the adulteration of camellia oil fast, efficiently and economicly.At present, the methods to detect the adulteration of edible vegetable oil include physical and chemical test, chromatography, near-infrared reflectance spectroscopy and so on. The first method is limited in detection range, the second one which needs expensive apparatus and professional laboratory personnel and is complex in the process of detection, makes it not suitable for site rapid testing. As a new non-destructive rapid detection technology, the third one has a broad prospect.The instrument uses near infrared light-emitting diodes combined with the corresponding filters to produce narrow-band and different center wavelength near-infrared light, uses OPT101as light intensity detecting sensor. The microcontroller capture the light intensity signal which is processed by signal conditioning and A/D convert circuit, and then send the data to PC by serial communication. On PC, we use MFC programming to receive the spectral data,and save all spectral data and the adulteration concentration of the sample set into a excel format file. Then we use Matlab programming to get all the data from the file and build up a mathematical model between the spectra data and the adulteration concentration by multiple linear regression analysis. Through this model we can predict the adulteration value of the unknown oil sample. In this paper, we detail hardware and software design of the instrument based on the concept of modular design. We prepare some oil samples,in which are different from the content of corn oil and camellia oil, and we gather these sample into a sample set. After that, we build up the model between spectral data and adutreation concentration of the sample set by multiple linear regression analysis. The result showed that:when corn oil adulteration concentration is of5%or more, multiple linear regression method can well predict the adulteration of the camellia oil, and the higher the rate of adulteration concentration, the higher the prediction accuracy.In this paper, we design a system for detecting the adulteration of camellia oil using near-infrared reflectance, and analyze the predictive ability of multiple linear regression method.This reaserch provide an effective mean for the protable fast and economic detection of adulteration of camellia oil.