Development of a CNN based robot vision system

Abstract

This final year project report presents the development and implementation of an object detection system using YOLOv5-obb, designed for detecting oriented bounding boxes (OBBs), and its integration with a UR5e (Universal Robots) robotic arm. The primary objective of this study is to detect cups using OBBs, label them according to their oriented positions, and manipulate the robotic arm to pick up a cup, move to another cup, and pour the contents accordingly.

The report begins with a comprehensive literature review, examining existing research on object detection, oriented bounding boxes, and YOLOv5-obb. The subsequent chapters outline the design and development process, including the implementation of the YOLOv5-obb algorithm, its integration with the UR5e robotic arm, and the testing procedures employed to evaluate the system's performance in detecting and manipulating cups.

The results chapter presents the findings from the experiments and testing conducted during the project. These results are analysed in the context of the project objectives, highlighting the effectiveness of the proposed system in accurately detecting cups with their oriented positions and controlling the robotic arm to perform the required tasks.

The report concludes with a summary of the project's achievements, a discussion of the limitations and challenges faced, and recommendations for future research and development in the field of object detection and robotic manipulation using oriented bounding boxes.