Feature-optimized convolutional modeling for predicting loquat soluble solids content from hyperspectral imaging with multi-algorithm wavelength selection

Key Points

• The analysis predicts soluble solids content effectively in loquats, showing high accuracy with various algorithms.
• A multi-algorithm wavelength selection process enhances the model's performance, identifying key spectral features.
• Convolutional modeling techniques applied to hyperspectral imaging provide detailed insights into fruit composition.
• Such modeling approaches may facilitate improved agricultural quality assessments, supporting better crop selection.