Research On Needle Tracking Based On Multi-sensor Fusion

With the rapid development of digital technology and surgical navigation technology,oral implant guidance technology has become the focus of modern oral implant navigation and related medical instruments.Aiming at the current hot issues in the field of planting navigation,this thesis studies the needle target tracking scheme based on multi-sensor fusion.VR positioning technology based on lighthouse laser and camera optical positioning technology is adopted.The fusion scheme proposed includes the pre-processing stage,the preliminary positioning solution stage based on different sensors,and the data fusion test stage.The research includes a target pose solution model based on different sensor positioning,error control of planting navigation,implementation and testing of multi-sensor fusion scheme,etc.,and the effectiveness of the fusion scheme is verified by measurement.The main research contents of this thesis include the following three parts:1)According to the theory of correlation localization and fusion algorithm,this thesis proposes a method of needle tracking based on multi-sensor fusion.This thesis selects hardware such as a lighthouse laser sensor,monocular camera,and surgical needle for oral implant surgery.The fusion localization scheme includes the pretreatment,initial localization calculation,and fusion stages.Firstly,an equipment calibration experiment is carried out,calibration parameters and errors are obtained,and then the location calculation model based on different sensors is introduced.2)The data fusion scheme of lighthouse laser and monocular camera is used to solve the problems of unstable initial positioning signal,limited detection range and increasing positioning error when the light source intensity changes.For the VR positioning part of the lighthouse laser,the data of laser scanning and signal return are used to solve the spatial attitude of the target.The initial results show that the stability of the data needs to be improved.According to the PNP algorithm,the camera part solves the target pose with a specific code.By setting the appropriate threshold of image frame processing,it is used as a sliding window to reduce the system computation and improve the real-time performance of the system.3)Improving strategies are proposed according to the advantages and disadvantages of different data fusion algorithms.Before fusion,the reference coordinate system of different positioning settlement models is jointly calibrated,and unified coordinates are completed.Due to the different working frequencies of different sensors,the two groups of data are downsampled and extracted based on the sampling time stamp.Finally,the system state model is established,and the improved EKF algorithm is adopted for multi-source data fusion processing.Then,the lighthouse laser’s unstable and hopping error data can be corrected by fusion data updating and monocular camera data.The simulation test shows that the tracking error can meet the accuracy requirements of the actual positioning error,and the calculation error can be controlled within0.18 mm.By comparing the positioning accuracy of 0.2mm of the current product in clinical implant navigation surgery,it is verified that the fusion positioning scheme proposed in this thesis has certain practical significance,and finally meets the surgical accuracy requirements of digital oral implant navigation.