MCD (Multi-Campus Dataset)
收藏资源简介:
MCD数据集是由南洋理工大学电气与电子工程学院等多个研究机构合作创建的,旨在为机器人感知提供一个多样化的大型多校区数据集。该数据集包含多种传感模式,如CCS和NRE激光雷达、高质量的IMU、摄像头和UWB传感器,并提供了高精度的地面实况数据。数据集覆盖了三个欧亚大学校区的多样化挑战性环境,包含18个序列,超过20万次激光雷达扫描和150万次摄像头帧。此外,数据集还包含了29个类别的语义标注,覆盖了59,000次稀疏NRE激光雷达扫描,为现有的语义分割研究提供了新的挑战。MCD数据集的应用领域广泛,旨在解决机器人感知中的多种问题,推动下一代成本效益高的机器人、人工智能和AR/VR系统的发展。
The MCD Dataset was collaboratively developed by multiple research institutions including the School of Electrical and Electronic Engineering of Nanyang Technological University, with the goal of providing a diverse, large-scale multi-campus dataset for robotic perception. This dataset integrates various sensing modalities, including CCS and NRE LiDARs, high-quality IMUs, cameras, and UWB sensors, and provides high-precision ground truth data. It covers diverse and challenging environments across three university campuses in Eurasia, comprising 18 sequences, over 200,000 LiDAR scans, and 1.5 million camera frames. Furthermore, the dataset includes semantic annotations for 29 categories, covering 59,000 sparse NRE LiDAR scans, which poses new challenges for existing semantic segmentation research. The MCD Dataset has a wide range of application scenarios, aiming to address various issues in robotic perception and advance the development of next-generation cost-effective robots, artificial intelligence, and AR/VR systems.



