The Analysis Of Tea Features Using ICP Spectrum

Tea, as the head of the three world drinks, has thousands of years of history and rich innutritional value and medicinal value. With the improvement of living standards, more andmore people enhance the concerns to food safety. As the hometown of tea, China has a longhistory of tea’s growth and many famous brand tea, and tea is widely distributed. To ensurethe quality and safety of tea, rapid identification of the type of tea, origin, picking time, andother characteristics have important practical significance for manufacturers and consumers. Itis necessary to apply ICP spectral analysis to study the properties of tea for it’s simple, fastand no pollution characteristics.This paper researched tea’s producing area and picking time by ICP spectrum data, usingthe method of feature extraction. The paper’s main conclusions are as follows:This paper not only carried out a detailed mathematical derivation to characteristic basistheory, but also describes the mathematical basis for its significance in geometry andstatistical.Use the methods of feature extracted twice of ICP spectral data which was collected andpretreated to identify the origin feature. It also analyzed the kinds of tea and element typesinvolved in. This paper selected15kinds of tea samples, which came from Hunan, AnhuiKeemun, Huangshan in Anhui, Fujian and Yaan in Sichuan, and10kinds of elements. Aftertwo times of features extraction and the establishment of feature-based,24kinds of tea will beprojected on the above-mentioned feature-based. The result show that: This method caneffectively identify the origin of tea.Use the methods of feature extracted twice of ICP spectral data to identify the pickingtime feature. For four kinds of tea samples, which picking time is several days before andafter Qing Ming day and picking standard is one leaf and a bud, and another four kinds of teasamples, which picking time is several days before and after Beginning of Summer andpicking standard is2-4leaves, this paper make two times of features extraction and create afeature-based. Then,18kinds of tea, which have various picking time and picking standard,will be projected on the above-mentioned feature-based. The result show that: This methodcan effectively identify the picking time of tea.