Abstract:Environmental and instrumental factors can cause variations in sample temperature, which may affect the predictive performance of models that use online Vis/NIR spectroscopy to predict the soluble solids content (SSC) of fresh fruit, an important quality indicator that influences their price. This study proposed a knowledge-guided (KG) temperature correction method based on one-dimensional convolutional neural network (1D-CNN) to reduce the impact of sample temperature on the SSC prediction model. First, a 1D-CNN was developed for prediction of temperature from spectra, and gradient-weighted class activation mapping was used to identify features related to temperature. These features, which were sensitive to temperature variations, were used to suppress temperature-sensitive bands and generated new spectra that were less affected by temperature. Subsequently, these new spectra were utilized to construct the partial least squares (PLS) model to predict SSC. To validate the efficacy of the proposed method, it was compared with the external parameter orthogonal method (EPO), competitive adaptive reweighted sampling (CARS) and successive projection algorithm (SPA) in terms of SSC prediction performance. The results indicated that while KG-PLS does not enhance the performance of local models on local (a specific temperature) testing sets, it can improve their performance on external (temperatures other than the specific temperature) testing sets. The root mean square error of cross validation (RMSECV) for SSC of local model at 15 degree celsius using KG-PLS was 0.415 %, which was 0.196 %, 0.166 %, 0.169 %, and 0.065 % lower than the RMSECV obtained by PLS (0.611 %), CARS-PLS (0.581 %), SPA-PLS (0.584 %), and EPO-PLS (0.480 %), respectively. In general, the proposed KG method can effectively mitigate the impact of temperature interference on spectral modeling, providing a reference for high-precision online non-destructive sorting of fruit quality.

Influence Of Temperature On Visible And Near-Infrared Spectra And The Predictive Ability Of Multivariate Models

Assessing the temperature influence on the soluble solids content of watermelon juice as measured by visible and near-infrared spectroscopy and chemometrics

Maintaining the Predictive Abilities of Tea Moisture Content Models on Different Environment Based on Visible and Near-Infrared Spectroscopy Technique

Study on the Influence of Temperature on Near Infrared Spectra

Broadening the Scope of Food Quality Control: a Calibration Transfer Strategy for Minimizing External Influences on Spectral Analysis

Comparison of Detection Modes in Terms of the Necessity of Visible Region (VIS) and Influence of the Peel on Soluble Solids Content (SSC) Determination of Navel Orange Using VIS–SWNIR Spectroscopy

Improving SSC Detection Accuracy of Cherry Tomatoes by Feature Synergy and Complementary Spectral Bands Combination

Analysis of Multi-Factor on Measurement Improvement of an Infrared Imager in Low-Temperature Environments

Improving Discrimination Accuracy of Pest-Infested Crabapples Using Vis/NIR Spectral Morphological Features

Vis/NIR spectroscopy technique for determination quality attributes of tomato fruit

Variable Selection in Visible and Near-Infrared Spectra: Application to On-Line Determination of Sugar Content in Pears

Visible/Near-Infrared Spectra for Linear and Nonlinear Calibrations: A Case to Predict Soluble Solids Contents and pH Value in Peach

Temperature Influence For Fourier Transform Near-Infrared Transmittance Measurement Of Citrus Fruit Soluble Solids Contents

Using Knowledge-Guided Temperature Correction for Online Non-Destructive Detection of Soluble Solids Content in Pear Via Vis/NIR Spectroscopy

Temperature Compensation for VIS/NIR Spectroscopy Measurement of Chinese Rice Wine Quality

Effect of Biological Variability on the Robustness of FT-NIR Models for Sugar Content of Pears

The Near Infrared Spectral Bands Optimal Selection in the Application of Liquor Fermented Grains Composition Analysis]

Development of Multi-Cultivar Models for Predicting the Soluble Solid Content and Firmness of European Pear (pyrus Communis L.) Using Portable Vis–nir Spectroscopy

Comparison of Chemometrics Methods for Assessing Internal Quality of Pears On-Line Using Visible/near Infrared Transmission Technique

Study on multi algorithms for modeling of NIR spectra and MT-firmness of pears

Effect of Fruit Moving Speed on Predicting Soluble Solids Content of ‘cuiguan’ Pears (pomaceae Pyrifolia Nakai Cv. Cuiguan) Using PLS and LS-SVM Regression