Description

This thesis focuses on evaluating and benchmarking state-of-the-art methods for object pose tracking in robotic systems, with a primary emphasis on a solution utilizing Nvidia FoundationPose in combination with an active stereo camera. The performance of this approach will be compared against other leading object pose estimation methods in terms of accuracy, robustness, and efficiency in dynamic and real-world scenarios. By analyzing the strengths and limitations of these methods, the research aims to provide insights that support the optimization and application of advanced object pose tracking solutions in industrial robotics.