| Abstract |
| Abstract—The performance and automatic-evaluation- system in laparoscopy surgery require 3D instrument tracking and reconstruction. Traditionally, a camera is a standard device for tracking and reconstruction. The hard conditions inside a patient’s body as well as the movement of patients make camera system to work non exactly. For this reason, an inertial measurement unit (IMU) which could be installed outside patient’s body to avoid damaging organ was newly integrated with surgery system. Same as camera sensor, IMU have ability to track movement directly. However, the IMU measurement can’t support a high accuracy because of bias elements. Thus, we plan to fuse information from camera and IMU sensors to estimate the position, velocity and direction of the instrument. The proposal system uses an Extended Kalman Filter to integrate information from different sources, recoup the biases of IMU, and track the instrument in a unified framework. The results of the experiment show the effectiveness of our method compared with image-based and IMU-based methods and the ground true of the instrument are estimated by a system with four cameras. |