Root Canal Therapy Of Robot-assisted Research On Mixed Reality Of Registration And Tracking

According to the survey of the World Health Organization,dental caries affects more than 2.4 billion people in the world.As the most common and highest incidence oral disease in the world,dental caries is very easy to cause root canal complication inflammation and lead to many high complication diseases in oral cavity when the medical resources are insufficient or treatment is not timely,Root canal treatment is an effective means of clinical treatment for dental caries and related diseases.However,in the current medical environment of China,there are problems such as shortage of medical resources and imbalance of doctor-patient ratio,and there are also problems such as limited operation space in the oral environment and poor visualization environment which affect the effective information collection during the operation,which seriously affect the quality of root canal treatment.In order to solve the above problems,this paper proposed a scheme of robot-assisted root canal therapy with a three-dimensional(3D)model in a mixed reality environment to provide the key information needed during the procedure.For this scheme,this paper has carried out the following research.In this paper,the patient’s oral CT image is used for 3D model reconstruction.After obtaining the patient’s oral CT image,the characteristics of the CT image of the affected tooth are analyzed.Before 3D model reconstruction,an improved Gabor wavelet filtering technology was proposed to preprocess the CT image.By quantifying the imaginary part value of Gabor wavelet,simplifying the calculation to improve the denoising effect,and then optimizing the region growth algorithm,traversing the window to automatically find the optimal seed point to segment the patient’s oral CT image,and then using the equivalence surface extraction algorithm to reconstruct the 3D model of the patient’s teeth and root canal to obtain the 3D reconstruction model of the patient’s teeth and root canal.A real-time display platform for the patient’s root canal was built to realize the need for intraoperative visualization information.The Holo Lens2 head-mounted display device combined with Unity 3D and MRTK components was used to develop the real-time display platform in a mixed reality environment.The reconstructed 3D model of the patient’s tooth and root canal was placed in the mixed reality environment,and the reconstructed model was subjected to eye-tracking,gesture interaction and spatial mapping functionsso that the reconstructed virtual model can interact with the user in real time.After completing the construction of the real-time display platform,the principle research and calibration experiments were conducted for the calibration of the camera position parameters required for the subsequent registration and tracking methods in this paper,so as to pave the way for the subsequent work.After completing the construction of the real-time display platform,the tracking and registration method is researched.Considering the problems of the SLAM system that comes with Holo Lens2 in terms of the first frame position and model pose fixation,the researched of registration and tracking algorithms needs to be completed in steps.First,the 3D model is registered using QR code,the RGB image of Holo Lens2 camera is acquired,the image is pre-processed with filtering and binarization,the center pixel coordinates of QR code are calculated by edge detection and center point detection,the recognized QR code is geometrically corrected and decoded,the camera pose is solved by constructing Pn P matrix,and the pose is optimized using the optimized SVD algorithm.Then the 3D model tracking algorithm is studied,the SLAM system is built,the feature point detection algorithm is improved and optimized in the visual odometry,the LATCH feature descriptor is combined with the improved LBP feature descriptor to propose the LATCH/LBP feature descriptor,the LATCH/LBP feature descriptor combined with the FAST corner point detection algorithm improves the efficiency of feature point matching,the optimized and improved feature point extraction matching algorithm increases the number of matching points to provide an effective guarantee for camera motion estimation and back-end optimization.After completing the 3D model tracking registration algorithm,the trajectory path planning of robotic cavity preparation and pulp opening can be carried out according to the central point of the buccal crown of the 3D model and the position of the root canal opening,and a reasonable trajectory path for cavity preparation and pulp opening can be planned and interpolated on the premise of meeting the clinical requirements of root canal treatment.Finally,experimental verification was carried out,experimental materials were prepared to build the robot assisted root canal therapy experimental platform in mixed reality environment,and the images acquired by Holo Lens2 were matched with feature points and calculated with camera trajectory coordinates.The matching results and calculation results were compared with the results of the traditional method,and the error of camera trajectory coordinates under the two algorithms were analyzed to verify the superiority of the improved algorithm in terms of the correct image matching rate.The error between the discrete point coordinates and the theoretical discrete point coordinates in the experimental robotic cavity preparation trajectory was analyzed,and the experiments showed that the robotic cavity preparation trajectory is feasible and the 3D model tracked and registered in the mixed reality environment can provide effective information for the dentist during the operation.