Abstract:

Rice is one of the most important food crops. Identification of rice seed variety is currently in agricultural production, crop breeding and seed testing of important issues. As species identification difficult, each of the species made a mistake and poor purity causing huge economic losses. This paper presents an application of photoluminescence spectroscopy combined with chemometrics rapid, non-destructive identification of a new method of rice varieties. In order to achieve the rapid discrimination of the varieties of rice seed, University of Agriculture collected from the Yangtze to 5 different varieties of rice were 125 samples of 650-1000 nm band photoluminescence spectra collected, for each species All 25 samples of rice data by Beijing Optical Instrument Co. Ltd. production Zhuoli Han photoluminescence spectrometer. The spectral data compressed by principal component analysis a smaller number of new variables (principal component), of which the first seven principal components could explain 99.892% of the original spectral information. Therefore the first 7 principal components as BP neural network input, the value of different rice varieties as the output of BP neural networks, pattern recognition model for the establishment of rice varieties. Samples were randomly divided into model contains the set of 100 samples and 25 samples of the prediction set. The results showed that 5 random model prediction 100% correct rate, and principal component analysis has good capability of data compression. It is indicated that the model set up by the combination of principal component analysis and BP neural network in the present study is rapid in analysis and precise in pattern discrimination. It can be concluded that a new approach to distinguishing rice seed was offered.