Study Of Technique For Optical Tracking Surgical Instrument In Surgery Navigation

With the rapid development of computer technology, precise instruments and imagingtechniques, surgical operation is entering into an age of surgery navigation. It has the vitalsignificance to improve the accuracy of locating the target, optimize the path of surgery,reduce the surgical trauma and increase the success rate in surgical operation. Surgerynavigation system (SNS) consists of3D visualization equipment and real-time positionmeasuring device could calculate relative position and direction between surgical instrumentsand patient accurately. However, optical tracking system is the most popular position devicein SNS today. It builds the position device based on near infrared light(NIR) with the researchof how to build binocular vision measurement system and accomplishment of matchingsoftware. In this paper, we describe all components according their positional relationship,and propose the algorithm of surgical instruments registration, stereo matching andidentification automatically. All of these laid the foundation of the surgery navigation systemgreatly. The major contributions in this dissertation include:(1) Build optical tracking platform with corresponding software. This platform could beused in capturing images and registration, location and3D visualization for interesting objects.It aids to research of some algorithm such as calibration, registration and stereo matchingthrough images captured by the platform.(2) Calibration of NIR tracking system. Calibration of system is to determine the intrinsicand extrinsic parameters of two cameras for tracking objects. However, this system can notcapture traditional calibration board. An accurate method is adopted to design the board andcalibrate this system with the accurate result to achieve the purpose of accurate calibration.(3) Surgical instruments registration algorithm. It pivots surgical instruments in which tipkeeps in the same location to get the tip’s3D coordinates. Then, registration procedure iscomplete while obtaining the relationship between the lighting points of surgical instrumentsand their tips. Mean square error (MSE) of distances between the lighting points and tipscould be used as elaborating the accuracy and robustness with the analysis of3D coordinatesat different locations.(4) Stereo matching surgical instruments algorithm. To calculate the3D coordinates ofsurgical instruments, the correspondence of lighting points on these instruments should bedetermined. After the sub-pixel coordinates of these lighting points on left and right image areextracted automatically, the line equations of all lighting points are obtained according to theintrinsic and extrinsic parameters. Two lines produced by the same lighting point are intersected with each other, from which the3D coordinates of these lighting points arecomputed to satisfy the distance relationship among points on surgical instruments to realizestereo matching. In this process, the identification of surgical instruments could be donequickly and accurately as well as calculating their3D coordinates, thus it improve theefficiency of the recognition and tracking surgical instruments in surgery navigation.