Polarization-sensitive OCT-based pearl feature detection

Abstract

In this study, a detection system based on Polarization-Sensitive Optical Coherence Tomography (PS-OCT) using Mueller Matrix Optical Coherence Tomography (MM-OCT) was developed. By employing PS-OCT technology, the system was able to fully detect all sixteen elements of the Mueller matrix. By comparing the intensity element M00 among the sixteen elements of the Mueller matrix, the texture structure of the pearl layers could be observed. This allowed for differentiation between freshwater and saltwater pearls, identification of genuine and fake pearls, detection of internal flaws in pearls, and differentiation between nucleated and non-nucleated pearls. The study also involved the labeling of connected regions in binary images, where pixels within the same connected region were assigned the same label. The labeled images were displayed to facilitate more intuitive qualitative analysis, and quantitative analysis was performed using gray-level co-occurrence matrices. Subsequently, pearl layer pixels were extracted from multiple angles in the images, and the thickness of the pearl layer was calculated using the extracted pixels and axial resolution. Finally, detection and classification of unknown pearls were conducted, yielding results consistent with the actual outcomes. The measured thickness results after sectioning matched the calculated results, providing evidence for the feasibility of the experimental method proposed in this study.