RadIOCD

This repository introduces the RadIOCD (Radar-based Interior Object Classification Dataset), which contains sparse point cloud representations of interior objects, collected by subjects wearing a commercial off-the-shelf mmWave radar. RadIOCD includes the recording of 10 volunteers, aged between 25 and 50 years old. A total amount of 5 objects, with the participants moving towards them in 2 different environments were recorded. RadIoCD includes sparse 3D point cloud data, together with their doppler velocity provided by the mmWave radar. The files were stored in CSV format to ensure its reuse. The scope of RadIoCD is the availability of data for the recognition of objects solely recorded by the mmWave radar, to be used in applications were the vision-based classification is not robust (e.g, in search and rescue operation where there is smoke inside a building). Furthermore, we showcase that this dataset contains enough data to apply Machine Learning-based techniques, and ensure that it could generalize in different environments and "unseen" subjects.

本开源仓库介绍了RadIOCD（基于雷达的室内物体分类数据集，Radar-based Interior Object Classification Dataset），该数据集包含室内物体的稀疏点云表征，数据采集自佩戴商用现货毫米波雷达的受试者。数据集涵盖10名年龄介于25至50岁之间的志愿者的采集记录，研究人员在2种不同环境中，记录了受试者朝向5类不同物体移动的场景数据。该数据集包含稀疏三维点云数据，以及毫米波雷达采集得到的多普勒速度信息。所有文件均以CSV格式存储，以保障数据集的可复用性。本数据集的应用场景为仅依靠毫米波雷达采集的数据开展物体识别任务，适用于基于视觉的分类方案鲁棒性不足的场景（例如建筑物内存在烟雾的搜救作业）。此外，本研究证明该数据集拥有足够的数据量以支撑基于机器学习的技术应用，并可确保模型在不同环境与"unseen"受试者数据上具备泛化能力。

搜集汇总

背景与挑战

背景概述

RadIOCD是一个基于毫米波雷达的室内物体分类数据集，包含稀疏3D点云和多普勒速度数据，由10名志愿者在两种不同环境中对5种物体进行记录。该数据集以CSV格式存储，适用于机器学习应用，特别针对视觉分类不鲁棒的场景（如烟雾环境下的搜救行动）。