Cleaning Robot Trajectory Video Tracking Dataset

With the widespread use of intelligent cleaning devices, cleaning robots have become an important part of modern home and office environments. However, current cleaning technology still faces many challenges in route planning and obstacle avoidance in complex environments. Existing solutions, such as laser or ultrasonic-based navigation systems, have certain limitations in accuracy and real-time performance. This dataset aims to enhance the autonomous navigation and dynamic environment adaptability of robots through video tracking technology. Data collection was carried out using high-definition cameras in various complex household and commercial environments, covering different lighting conditions and floor materials. The data has undergone multiple rounds of expert annotation to ensure high precision and consistency in trajectory and obstacle annotations. The annotation team is composed of computer vision and robotics experts, with a team size of over 20 people. Preprocessing steps include video stabilization, frame extraction, data augmentation, and more. The data is organized by scene and complexity for easy retrieval and use.