A general data tool of mining graphic feature patterns from photonic crystal spheres

Photonic crystal sensing applications have been widely used in numerous domains. In this work, we present a comprehensive framework that leverages photonic crystal sensing through advanced image processing and machine learning techniques. Unlike many previous works, which only utilize color and spectrum information for photonic crystal sensing, this research aims to classify photonic crystal morphologies based on their pattern changes in the RGB camera image domain alone. By utilizing computer vision and machine learning methods for pattern extraction, topological analysis, dimensionality reduction, and neural feature analysis, our findings demonstrate that the observed patterns in photonic crystal sensing indeed correlate with chemical imprinting.