Medical tracking is a significant issue in vision-based robotic-assisted surgical navigation, especially for distal locking of intramedullary nails. Existing solutions face limitations such as high manufacturing costs for targets, complex tracking schemes, and low positioning1precision. This paper proposes a novel method to estimate the pose of medical targets through pre-calibration and the Perspective-n-Point (PnP) algorithm, which determines the position of the distal intramedullary nail hole and projects this position onto the monitor. The precision of medical target positioning is highly affected by the distortion coefficient of camera internal parameters. To address this, we design and construct a distortion compensation model to reduce its impact on positioning precision. Additionally, to mitigate the effect of illumination variations, automatic exposure of polarized vision is utilized. Through 50 reprojection experiments, the proposed distortion model achieves a positioning precision of 0.284 mm at a working distance of one meter, significantly outperforming the 0.4 mm precision of the division model and the 0.426 mm precision of the polynomial model with a relative improvement of 30% and 34.2%. This method enhances the accuracy and reliability of robot-assisted surgical navigation, facilitating more precise and efficient surgical procedures.

Owing to the rapid development of emerging 360$^{\circ }$panoramic imaging techniques, indoor 360$^{\circ }$depth estimation has aroused extensive attention in the community. Due to the lack of available ground truth depth data, it is extremely urgent to model indoor 360$^{\circ }$depth estimation in self-supervised mode. However, self-supervised 360$^{\circ }$depth estimation suffers from two major limitations. One is the distortion and network training problems caused by Equirectangular projection (ERP), and the other is that texture-less regions are quite difficult to back-propagate in self-supervised mode. Hence, to address the above issues, we introduce spherical view synthesis for learning self-supervised 360$^{\circ }$depth estimation. Specifically, to alleviate the ERP-related problems, we first propose a dual-branch distortion-aware network to produce the coarse depth map, including a distortion-aware module and a hybrid projection fusion module. Subsequently, the coarse depth map is utilized for spherical view synthesis, in which a spherically weighted loss function for view reconstruction and depth smoothing is investigated to optimize the projection distribution problem of 360$^{\circ }$images. In addition, two structural regularities of indoor 360$^{\circ }$scenes are devised as two additional supervisory signals to efficiently optimize our self-supervised 360$^{\circ }$depth estimation model, containing the principal-direction normal constraint and the co-planar depth constraint. The principal-direction normal constraint is designed to align the normal of the 360$^{\circ }$image with the direction of the vanishing points. Meanwhile, we employ the co-planar depth constraint to fit the estimated depth of each pixel through its 3D plane. Finally, a depth map is obtained for the 360$^{\circ }$image. Experimental results illustrate that our proposed method achieves superior performance than the current advanced depth estimation methods on four publicly available datasets.