Method for selecting subinterval of near-infrared spectrum wavelength based on simulated annealing algorithm

Abstract

The invention claims a method for selecting a subinterval of near-infrared spectrum wavelength based on a simulated annealing algorithm for analyzing the quality of farm products or foods, which comprises the following steps: performing pretreatment on the near-infrared spectrum and dividing the full spectrum of the near-infrared spectrum subjected to pretreatment into k subintervals; selecting the optimal number of characteristic subintervals and the combination mode thereof by using the simulated annealing algorithm, and then increasing the number of subintervals progressively; repeating the simulated annealing algorithm and then comparing the global optimal characteristic subinterval combinations obtained from the optimal characteristic subinterval combination in various interval division modes when finishing; and finally establishing a correction set model and a prediction set model. Apart from scientific and rational division of spectrum sub-intervals, the method has the advantages of ensuring the rapid convergence of algorithm, avoiding the problem that the algorithm only realizes the local optimal solution by judging the importance of the newly selected subinterval according to Metropolis rule, reducing the calculated amount for modeling, accelerating modeling and obtaining the high-quality near-infrared spectrum model quickly on the premise of obtaining the global optimal solution.

Description

technical field [0001] The invention relates to a method for selecting a near-infrared spectrum characteristic wavelength subinterval for analyzing the quality of agricultural products or food, in particular to a method for selecting a near-infrared spectrum characteristic wavelength subinterval based on a simulated annealing algorithm. Background technique [0002] At present, near-infrared spectroscopy is more and more widely used in the analysis of agricultural products and food quality. However, the near-infrared technology itself also has some shortcomings, such as complex background, low information intensity, overlapping peaks, etc., making it difficult to analyze near-infrared spectra with conventional spectral analysis methods. How to effectively extract features from a large number of near-infrared spectral data Information has become the focus of domestic and foreign scholars' research. The characteristic absorption of the sample in one or several bands of the ne...

AI Technical Summary

Problems solved by technology

Although these algorithms can extract the characteristic information of the spectrum, the process of dividing the sub-intervals has a certain degree of subjectivity. At the same time, the interval combination method is single, which cannot ensure the global optimal combination.

Images