Study on Hierarchical Threshold De-noising Method Based on near Infrared Spectrum Data

In recent years, the research of modeling method based on near infrared spectrum data has become one of the main methods for the analysis of mineral composition, however, due to the influence of various factors, there is a lot of noise in the near infrared spectrum data, which causes serious influence on the precision of the model and model robustness. In this paper, the method of using wavelet detail coefficients of autocorrelation for hierarchical threshold denoising is proposed to eliminate the noise contained in near infrared spectrum data of hematite. First, the maximum decomposition layer is determined according to the minimum frequency of effective signal. Second, the signal is decomposed to the maximum degree, and the threshold value is determined according to the correlation of the coefficients of each decomposition layer and the noise. Then, the near infrared spectrum data is processed by the calculated threshold value. Results show that the method not only can eliminate the noise in the data effectively but also can maximize the retention of feature information in the data, which improve the precision and robustness of the model effectively, and an effective de-noising method was provided for the models establishment based on near infrared spectrum data of hematite.