This study presents a comprehensive multi-sensor dataset designed for 3D mapping in challenging indoor and outdoor environments. The dataset comprises data from infrared cameras, depth cameras, LiDAR, and 4D millimeter-wave radar, facilitating exploration of advanced perception and mapping techniques. Integration of diverse sensor data enhances perceptual capabilities in extreme conditions such as rain, snow, and uneven road surfaces. The dataset also includes interactive robot data at different speeds indoors and outdoors, providing a realistic background environment. Slam comparisons between similar routes are conducted, analyzing the influence of different complex scenes on various sensors. Various SLAM algorithms are employed to process the dataset, revealing performance differences among algorithms in different scenarios. In summary, this dataset addresses the problem of data scarcity in special environments, fostering the development of perception and mapping algorithms for extreme conditions. Leveraging multi-sensor data including infrared, depth cameras, LiDAR, 4D millimeter-wave radar, and robot interactions, the dataset advances intelligent mapping and perception capabilities.Our dataset is available at https://github.com/GongWeiSheng/DIDLM.

翻译：本研究提出一个面向室内外复杂环境3D映射的综合多传感器数据集。该数据集包含红外相机、深度相机、激光雷达与4D毫米波雷达等多源数据，为探索先进感知与映射技术提供支撑。多传感器数据的融合增强了在雨雪、凹凸路面等极端条件下的感知能力。数据集还收录了不同速度下室内外机器人的交互数据，提供了真实的背景环境。通过对比相似路径的SLAM结果，分析了不同复杂场景对各传感器的影响。采用多种SLAM算法处理该数据集，揭示了不同算法在不同场景下的性能差异。总之，本数据集解决了特殊环境数据稀缺的问题，促进了极端条件下感知与映射算法的发展。依托包括红外、深度相机、激光雷达、4D毫米波雷达及机器人交互在内的多传感器数据，该数据集推动了智能映射与感知能力的提升。数据集可通过 https://github.com/GongWeiSheng/DIDLM 获取。