Real-Time Object Tracking and Pose Prediction for Robotic Manipulation using ROS

This thesis explores the use of an eye-in-hand camera system, where the camera is mounted on the end-effector of a robotic arm, to perform a pick-and-place routine. The objective is to develop techniques that enable the robotic arm to pick objects from a conveyor belt efficiently. The key steps involved in this process include camera calibration, object detection and tracking, as well as the implementation of an object following and picking algorithm. Each of these steps is addressed in detail, providing a comprehensive approach to robotic manipulation using this camera system.