Near-infrared spectrum feature extraction method and device

Abstract

The invention discloses a near-infrared spectrum feature extraction method and device. The method comprises the following steps: acquiring to-be-detected N samples; acquiring near-infrared spectrum data of the N to-be-detected samples by using a spectrograph; preprocessing the near-infrared spectrum data to obtain two-dimensional near-infrared spectrum smooth data; arranging arrangement and conversion on the two-dimensional near-infrared spectrum smooth data to obtain four-dimensional spectrogram data; carrying out feature extraction on the four-dimensional spectrogram data; and then carryingout feature arrangement on the four-dimensional spectrogram data after feature extraction to obtain two-dimensional feature data. The near-infrared spectrum feature extraction method and device have the following advantages: the data integrity can be guaranteed; feature extraction can be carried out in a full-spectrum interval; and the information is protected from being lost.

Description

technical field [0001] The invention relates to the fields of pattern recognition and non-destructive testing, and more specifically relates to a near-infrared spectrum feature extraction method and device. Background technique [0002] Near-infrared spectroscopy analysis technology is an analysis method that utilizes the optical characteristics of chemical substances in the near-infrared spectrum range to achieve qualitative and quantitative rapid detection of detection objects. It has the advantages of less sample consumption, no sample damage, fast analysis speed, and low detection cost. , No waste pollution and other conventional detection and analysis methods can not match the advantages. After years of technological development and improvement, this technology has been widely used in important national production fields such as agriculture, petroleum, medicine, chemical industry, and food. With the continuous development of my country's market economy and the improvem...

AI Technical Summary

Problems solved by technology

The existing feature extraction algorithms have different effects according to the characteristics of the analysis object and the acquired data. Specifically, all analysis methods are not universal and can only work on the analysis objects with one or several data structures. When detecting When the frequency of object changes is high, the effectiveness of existing analysis methods cannot be ensured;

[0004] (2) The integrity of feature data is low

The integrity of the characteristic data determines the validity, stability and comprehension of the built model. The existing analysis methods can only realize the selection or data compression of the near-spectral data interval, but cannot realize the feature extraction of the full-spectral interval, so the final analysis method cannot be guaranteed. The integrity of modeling data makes it difficult to optimize existing analysis models

[0005] (3) There are limitations in the feature extraction results

The invention solves the problem that the model structure does not have universality caused by excessively high or uncertain data dimensions in processing spectral data

However, the input is a spectrum image, and the two-dimensional data sample cannot be analyzed, resulting in the loss of the two-dimensional data sample, the feature extraction of the full spectral range cannot be realized, and the integrity of the final data cannot be guaranteed.

Images