An Accuracy Improvement Method Based on Multi-Source Information Fusion and Deep Learning for TSSC and Water Content Nondestructive Detection in “Luogang” Orange

The objective of this study was to find an efficient method for measuring the total soluble solid content (TSSC) and water “Luogang” orange. Quick, accurate, nondestructive detection tools (VIS/NIR spectroscopy, NIR machine vision, electronic nose), four data processing methods (Savitzky–Golay (SG), genetic algorithm (GA), multi-source information fusion (MIF), convolutional neural network (CNN) as deep learning method, a partial least squares regression (PLSR) modeling method) were compared investigated. results showed that optimal TSSC based on VIS/NIR vision by SG + GA CNN PLSR. R2 RMSE 0.8580 0.4276, respectively. result 0.7013 0.0063, This optimized largely improved internal quality accuracy orange when from single tool with traditional provides reference improvement other fruits.