Image processing algorithms for real-time tracking and control of an active catheter

In this paper, we investigate vision-based robot-assisted active catheter insertion. A map of the vessels is extracted using image processing techniques and the locations of the junctions of the blood vessels are detected. The desired path of the catheter and the target is selected by the user/clinician. The tip of the catheter is tracked in real-time and the robot and the active catheter are controlled based on the position of the catheter inside the vessels. The active catheter is commanded by an autonomous guidance algorithm to bend in the appropriate direction at the branches. The stroke length for the robotic insertion is controlled by the autonomous guidance algorithm to ensure smooth motion of the catheter inside arteries. A PI controller has been implemented to overcome flexing in the catheter and maintain smooth motion. The catheter is autonomously guided from the point of entry to the target via appropriate commands, thereby shielding the surgeon from radiation exposure due to the X-rays in X-ray fluoroscopy and relieving him/her of stress and fatigue. Experimental results for the insertion algorithms are shown using a laboratory testbed.