Non-Destructive Prediction of Fruit Ripeness and Firmness Using Hyperspectral Imaging and Lightweight Machine Learning Models
For agricultural post-harvest quality assessment, this work provides a practical, low-cost alternative to expensive hyperspectral cameras and deep learning models, though the approach is incremental in applying classical ML to a known problem.
This study shows that lightweight tree-based machine learning models on hyperspectral imaging data can outperform deep learning models for fruit ripeness and firmness prediction, achieving over 94% of full-spectrum accuracy using only three visible-range wavelengths, enabling low-cost multispectral sensors as practical alternatives.
Post-harvest fruit quality assessment is essential for reducing food waste, yet reliable non-destructive methods typically depend on expensive hyperspectral cameras and computationally intensive deep learning models. These systems typically require GPU resources, large-scale training data, and domain expertise, limiting their feasibility for many real-world agricultural settings. This study systematically evaluates 20 classical machine learning algorithms on hyperspectral imaging data for simultaneous ripeness classification and firmness prediction across five fruit species, using cross-validated experimental design with Bayesian hyperparameter optimization. Data preprocessing strategy, particularly class balancing and spectral transformations, contributes as much to prediction accuracy as algorithm choice. Our results show that tree-based machine learning models can outperform state-of-the-art deep earning models reported in Fruit-HSNet. Moreover, the findings indicate that only three visible-range wavelengths are needed to recover over 94% of full-spectrum accuracy, demonstrating that low-cost multispectral sensors combined with lightweight machine learning models can serve as practical alternatives to expensive hyperspectral cameras and complex deep learning approaches for practical fruit quality sorting.