ISBN 9783843928458

978-3-8439-2845-8, Reihe Informatik

Philip Nicolai
A 3D Camera-based System Concept for Safe and Intuitive Use of a Surgical Robot System

185 Seiten, Dissertation Karlsruher Institut für Technologie (2016), Hardcover, B5

Zusammenfassung / Abstract

In today's robot-assisted interventions, the medical personnel has to devote a significant part of their attention to the surgical robot system to ensure and monitor its seamless functioning. This thesis explores the use of 3D cameras for automatic supervision of surgical robot systems, thereby contributing to their safe use and allowing the medical personnel to fully focus on their medical task.

An extensive overview of the state of the art serves as an introduction to relevant topics, such as surgical robot systems, range imaging cameras and their medical use, and safe human-robot interaction. Based on these findings, an Operating Room monitoring system is proposed that provides real-time detection of potentially harmful situations between the robot, its surroundings and the persons in its vicinity, and employs Spatial Augmented Reality for providing feedback to the personnel. The concept is realized and extensively evaluated, paying close attention to crucial practical considerations such as the effects of sterile surgical drapings on scene acquisition by 3D cameras.

Results show that the novel Shape Cropping algorithm can successfully detect the robot's pose and serve to establish a virtual safety zone around the robot. This allows for early detection of impending collisions, independent of the presence of surgical draping. The proposed algorithm for forward propagation of semantic labelling is applied to human tracking and shown to effectively bridge gaps in tracking between camera subsystems with different timing characteristics.