Research On Tracking Method In Computer Assisted Surgery Navigation System

With the increasing employment of the binocular stereo vision technology in computeraided operation navigation (abbr. operation navigation), the real-time and precision of brainsurgery will be enhanced greatly. The direct tool for doctors to conduct a operation is theoperation device, so whether it can be tracked and positioned real-timely and accurately isthe key to the success of the operation navigation.In order to track the device precisely during the operation, it is necessary to calibrate thecomputer aided operation navigation system (abbr. The operation navigation system) beforesurgery. This paper focuses on the binocular camera calibration, marker detection andextraction, target tracking algorithm and its critical techniques, involved in the tracking ofthe operation devices with the operation navigation system. Therefore, the real-time andprecision for tracking the working points of the equipment during the operation can beachieved eventually. The details of this study are as follows:1. Tracking system is designed. After studying the principle of binocular stereo visionand the camera imaging model, Halcon as a software is adopted to calibrate the binocularcamera, based on the method of camera calibration, taking the real environment of trackingdevices into consideration. The mapping, established between image coordinates and worldcoordinates with Halcon, combines nonlinear factors, which are generally ignored by thetraditional calibration method, so as to improve the accuracy of the camera calibration.According to the experiments, this means can meet the limitation demands of the calibrationresult error of the tracking system.2. The camera calibration template is set up. The calibration template with industrystandard precision is built into the software, Halcon, which solves the problem of lowprecision previously. Active light-emitting operation equipment used in the tracking systemis originally devised and produced. The overall layout of the additional markers on thedevices has been made, in accordance with the surgery environment and the features of thedevices being tracked.3. The study on the method of how to detect and extract the additional markers onoperation devices is conducted. The Otsy algorithm with no requirement of a prioriknowledge has obvious segmentation effect on the bimodal histogram. Furthermore, thecalibration of the operation devices is achieved with the establishment of their correspondingcoordinates, based on the principle and the technique of operation device calibration. As theexperiments indicate, the accuracy of the algorithm can meet the need of the tracking system,and the experiment achieves the desired effect of calibration.4. The comparison of Meanshift tracking algorithm and Camshift tracking algorithm is studied. In a non-uniform motion state, tracking the mark points of the operation tool withthe Meanshift algorithm leads to larger deviation. Additionally, due to the lack of thetemplate update mechanism, this algorithm can’t promise the effective tracking when thetarget turns deformed. Therefore, Meanshift algorithm, based on the Hue component of HSVcolor space, is applied to boost the accuracy and the robustness of the tracking system.Experiments prove that, in terms of tracking effect, Camshift algorithm outweighs Meanshiftalgorithm under the same environment.