Apple acidity near-infrared nondestructive testing method based on fusion characteristic wavelength selection algorithm

Abstract

The invention discloses an apple acidity near-infrared nondestructive testing method based on a fusion characteristic wavelength selection algorithm. The method comprises the following steps of: 1, acquiring spectral information of an apple sample mark point region, and measuring acidity data of the apple sample mark point region; 2, preprocessing the acquired spectrum; 3, performing characteristic wavelength selection by using a successive projection algorithm SPA and a competitive adaptive reweighted sampling algorithm CARS respectively, and fusing the characteristic wavelengths selected bythe SPA and the CARS; and 4, establishing a partial least squares (PLS) prediction model of apple acidity on the calibration set according to the spectrum and acidity data corresponding to the fused characteristic wavelength, and evaluating a model result on the prediction set. According to the method, the number of the selected wavelength variables and the accuracy of the established model can beconsidered, the model is simple, the detection efficiency is high, the practicability is high, an important means can be provided for rapid nondestructive detection of the acidity of apples, and thedevelopment cost of portable near-infrared instruments for specific purposes is reduced.

Description

technical field [0001] The embodiment of the present invention relates to the technical field of apple quality detection, in particular to a near-infrared non-destructive detection method for acidity of apples based on a fusion feature wavelength selection algorithm. Background technique [0002] Acidity is one of the important indicators to measure the internal quality of apples, which affects consumers' willingness to buy. Traditional apple acidity measurement methods are destructive and time-consuming. Near-infrared spectroscopy has been widely used in the internal quality inspection of apples due to its fast and non-destructive advantages, but it is rarely used in the detection of acidity. [0003] In the process of establishing a non-destructive testing model based on near-infrared spectroscopy, the selection of wavelength variables is an important part. Screening the characteristic wavelength or wavelength interval by a specific method can simplify the model on the o...

AI Technical Summary

Problems solved by technology

The number of wavelength variables selected by the SPA algorithm is small, but the accuracy of the final model is not good; the accuracy of the model finally established after the characteristic wavelength selection by the CARS algorithm is high, but the number of selected wavelength variables is also large; if Using the two together to form the CARS-SPA algorithm, the number of wavelength variables can be effectively reduced, but the accuracy of the model is also greatly reduced

In the field of non-destructive testing of apple acidity, the current single or combined characteristic wavelength selection algorithm cannot take into account the accuracy of the model and the number of wavelength variables, and there are problems such as low prediction accuracy, complex models, low detection efficiency, and weak practicability. Relatively high development costs for portable NIR instruments

Images