A visual tracking-based algorithm for real-time probe localization and 3D reconstruction in optical coherence tomography

Optical coherence tomography (OCT) represents an advanced, real-time, and non-invasive imaging modality that has found widespread application in both biomedical diagnostics and materials inspection. Despite its versatility, the absence of global positional feedback for the OCT probe constrains the system’s ability to conduct large-area scans, thereby limiting its applicability across a wider range of contexts. In this work, we introduce a novel visual tracking algorithm designed for real-time localization of the OCT probe. By integrating spatial position data with captured OCT images, this approach facilitates the potential for three-dimensional (3D) image reconstruction. The proposed method employs an AprilTag-based tracking framework, selected for its robustness in dynamic environments. Fiducial markers (AprilTags) are utilized to enhance the accuracy and reliability of probe localization across six degrees of freedom (6-DOF). Using a high-precision motion stage, we evaluated positional errors in both two-dimensional (2D) and three-dimensional (3D) scanning scenarios. The results demonstrate that our system achieves localization accuracy with an average position error as low as 2.44 × 10⁻⁵ <i>m</i> at the processing speed 83 frames per second. Compared to existing OCT localization techniques, our method offers a precise, cost-effective, and portable solution for large-area probe localization.