AR-Based Surgical Navigation System Based on Dual Plane Calibration and 2D/3D Registration

Surgical procedures are often associated with a range of potential risks, including nerve tissue damage resulting from complex lesion structures, narrow surgical spaces, or the low accuracy of the operating surgeon. The rise of augmented reality technology has attracted wide attention in the field of complementary medicine. The safe and low delay augmented-reality assisted surgery system can effectively guide the accurate operation. At the same time, augmented reality technology can make surgical sites more intuitive. In this work, we propose a novel surgical operating system that leverages coordinate system transformation and optical navigation to improve precision and safety. Our approach involves the integration of medical images and optical tracking devices before and during surgery to achieve coordinate transformation between the patient’s position, surgical instruments, and the pre-surgery CT scan. This process involves calibration, registration, and other critical procedures. At the same time, the study also introduced the function of augmented reality. Doctors wearing augmented reality glasses can use the system to see more vivid scenes of surgery. During the calibration phase, we utilize biplanar and interpolation algorithms to obtain the conversion relationship between the C-arm emission source and the X-ray image coordinate system. In the registration stage, we convert the preoperative 3D CT into a DRR image using digital reconstruction image technology, which we then register with the 2D X-ray during the operation. To optimize the iteration, we employ the improved Powell algorithm. The algorithm demonstrates an average angle and position error of 0.62° and 0.56mm, respectively. Through the transformation of spatial position relationship, the transformation relationship between preoperative CT and intraoperative patient position was obtained. A user-friendly, integrated software system has been developed that allows users to view the relative relationship between surgical instruments and patient position in CT in real time during surgery. An augmented reality module has also been introduced, allowing operators to see specific surgical models in specific glasses. In this study, the establishment and model test of the algorithm were realized, thus providing auxiliary functions for high precision surgical operation.