Tea leaf infrared spectrum classification method of fuzzy Kohonen identification clustering network

Abstract

The invention discloses a tea leaf infrared spectrum classification method of a fuzzy Kohonen identification clustering network. The tea leaf infrared spectrum classification method comprises the following steps: step 1, collecting a tea leaf sample infrared spectrum; step 2, carrying out multiplicative scatter correction (MSC) of the tea leaf sample infrared spectrum; step 3, carrying out dimension reduction processing on the infrared spectrum of a tea leaf by utilizing a main component analysis method and further carrying out characteristic extraction and dimension reduction by utilizing a linear discrimination method; step 4, carrying out fuzzy C-mean clustering to obtain an initial clustering center; step 5, classifying varieties of the tea leaves by adopting a fuzzy Kohonen identification clustering network method. According to the tea leaf infrared spectrum classification method disclosed by the invention, the problem that identification information of data cannot be extracted in a process of carrying out fuzzy clustering on the data by the fuzzy Kohonen clustering network so that the clustering accuracy is not high is solved. The identification information of tea leaf spectrum data can be dynamically extracted in a clustering process; the tea leaf infrared spectrum classification method has a rapid detection speed, high classification accuracy and high classification efficiency and can be used for classifying the tea leaves with the different varieties.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, in particular to a tea infrared spectrum classification method based on fuzzy Kohonen discrimination clustering network. Background technique [0002] Tea is one of the main crops in my country. Post-harvest treatment, quality judgment and testing of tea have always been an important means of tea quality assurance. At present, due to the lack of effective tea identification methods in China's tea market, the phenomenon of OEM in the tea market, shoddy and false ones are more serious, so the identification of tea varieties has become more and more important, and research on a simple and fast method The identification method of tea variety is very necessary. [0003] Infrared spectroscopy has the advantages of fast detection speed and simultaneous detection of multiple components. Different varieties of tea often have different components and contents, so the diffuse reflectance s...

AI Technical Summary

Problems solved by technology

As an unsupervised clustering method, the fuzzy Kohonen clustering network can only achieve fuzzy clustering of data, but cannot extract the identification information of data in the process of fuzzy clustering, so that high clustering accuracy cannot be obtained.

Images