Random Convolutional Kernel Transformation and Regression Models for Chlorophyll Estimation in Leaves from Reflection Spectra

This study compares traditional and advanced machine learning methods, such as MiniRocket and Deep Neural Networks, for non-destructive estimation of chlorophyll content in plant leaves using reflection spectroscopy within the 400–800 nm spectral range. The results show that MiniRocket offers the highest accuracy, with an $R^{2}$ of 0.982 (test data), highlighting its effectiveness in providing accurate chlorophyll predictions.