In this study, a 900-1700nm hyperspectral camera was applied, and FS-15, the product of Hangzhou Color Spectrum Technology Co., LTD., could be used for related research. Short-wave near-infrared hyperspectral camera, the acquisition speed of the full spectrum up to 200FPS, is widely used in the composition identification, substance identification, machine vision, agricultural product quality, screen detection and other fields.
Maple scoop is the initial processing product of medicinal Dendrobium. The production process is to cut off some of the roots of the fresh stem of Dendrobium, heat it while twisting it into a spiral shape or spring shape, and dry it into a concave. Maple bucket is a famous tonic in our country, mainly including tin maple bucket and purple maple bucket and other types. Among them, tin maple bucket with its good Yin and blood, throat protection effect, has become a kind of high-end health tonic, in East China, South China, especially in Hong Kong, Taiwan, foreign countries in Japan, Southeast Asia, is generally considered to be a high-grade health care products, the market demand is large. Dendrobium officinale is the only source of processing dendrobium officinale collected in China's pharmacopeia, the wild resources are very limited, although there is a certain scale of cultivation, but can not meet the current market demand, so the price is expensive, but also caused a lot of other dendrobium processing of Dendrobium officinale confused products. Such as dendrobium hairpin, dendrobium dentate (commonly known as purple skin Dendrobium) and other raw materials processing of the so-called "iron maple bucket". The medicinal value of Dendrobium officinale from different sources is very different, which seriously affects the stability of commodity quality, but also damages the rights of consumers, so it is very meaningful to establish an effective identification method for Dendrobium officinale. In addition, the price and medicinal value of Dendrobium officinale of different origin vary greatly. In the face of such a chaotic market, it is very necessary to study a rapid identification technology for Dendrobium officinale origin.
In this paper, the identification of three easily confused Dendrobium species was studied by using a miniature near infrared spectrometer, and an Internet of Things identification system for Dendrobium was constructed.
(1) This study investigated the influence of different sample sizes on NIR spectra and the establishment and evaluation of SIMCA models of different sample sizes. The results showed that with the increase of sample size, the larger the sample reflectance spectrum (log1/R), the higher the absorption intensity, and the NIR absorption intensity was positively correlated with the sample particle size. The model established by 80-eye spectrum is better than the model established by 40-eye and 60-eye spectrum.
(2) The near infrared diffuse reflectance spectra of intact and crushed maple bucket samples were quickly identified by SIMCA modeling. The results showed that after different pretreatment methods: The best preprocessing method for the complete sample spectrum is NAS plus S-G smoothing plus 1 order S-G derivative, and the best preprocessing method for the crushed sample spectrum is S-G smoothing plus 1 order S-G derivative plus mean centralization. The correct recognition rate and rejection rate of the models established for both the complete sample and the crushed sample are 100%, and the prediction stability and accuracy are high. Can realize the rapid identification of three kinds of maple bucket, more suitable for practical application.
(3) Using the best model built by the spectra of Fengdou complete samples and crushed samples to identify other samples, the results show that the rejection rate of all models is 100%, and the reliability of the model is good.