Research On The Algorithm Of Multi-objective Tracking In Surgical Navigation System

Research on the algorithm of multi-objective tracking in surgical navigation system isthe most important for improving surgical efficiency, increase the positioning of the surgicalaccuracy and reduce the risk of surgery. In recent years, object tracking technology has beengreatly developed, but there are still some problems in applications, the limitations of thesingle target tracking; impact of the environment an the shadows; the requirements of thespeed and accuracy, and so on, these problems can all affect the results of the tracking.The work and the result of research are as follows:1、In order to remove the shadow, this paper designed the blue tracking-ball, andproposed the methods of shadow detection in HSV color space. Select the appropriatecomponent, so the component are processed, and the coordinates of the centroid areextracted. The projection light as an important part of tracking and feedback, projection lightcorner were detected through the method of Harris corner detection which was improved.2、In order to improve the limitations of the single target in tracking process, a two-steptracking strategy were proposed. After the target locating and rough tracking by the methodof background subtraction and Mean Shift, accurate tracking can be achieved. And feedbackthe position to the computer for the next step through the projection light.3、According to the coordinates of the centroid of the target, the trajectory of the targetcan be predicted by the method of the least-squares fitting, and provide the basis for thetracking.The experiments show that the proposed method for the detection of the effect of theshadow is obvious, the centroid and the coordinates of projector light corner were accurate,and the accuracy were less than1mm which was comply with the requirement for accuratetracking. This study provides a good basis for surgical navigation system.