{
  "paper_id": "efd0123303cf398a0624526d69ef16ba",
  "paper_title": "InLiDa: A 3D Lidar Dataset for People Detection and Tracking in Indoor Environments",
  "paper_authors": [
    "Cristina Romero-González",
    "Álvaro Villena",
    "Daniel González-Medina",
    "Jesus Martínez-Gómez",
    "Luis Rodríguez-Ruiz",
    "Ismael García-Varea"
  ],
  "publication_date": "",
  "publication_venue": "International Conference on Computer Vision Theory and Applications",
  "abstract": "The objective evaluation of people detectors and trackers is essential to develop high performance and general purpose solutions to these problems. This evaluation can be easily done thanks to the use of annotated datasets, but there are some combinations of sensors and scopes that have not been extensively explored. Namely, the application of large range 3D sensors in indoor environments for people detection purposes has been sparsely studied. To fill this gap, we propose InLiDa, a dataset that consists of six different sequences acquired in two different large indoor environments. The dataset is released with a set of tools valid for its use as benchmark for people detection and tracking proposals. Also baseline results obtained with state-of-the-art techniques for people detection and tracking are presented.",
  "keywords": [
    "Indoor Lidar Dataset",
    "People Detection",
    "People Tracking",
    "Benchmark"
  ],
  "paper_institutions": [
    "University of Castilla-La Mancha"
  ],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:47:26.944923Z",
  "template_tag": "simple",
  "summary": "### 摘要\n本文针对室内环境中大型3D激光雷达用于人员检测与跟踪的数据集稀缺问题，提出了InLiDa数据集。该数据集包含六个序列，分别采自一个学术建筑的走廊和大学校舍的大厅，使用Velodyne VLP-16激光雷达采集3D点云，并以点级标注了人和机器人的位置，同时为每个移动个体分配唯一标识以便跟踪。数据集中还包含20秒的无动态对象子序列用于背景建模，并提供rosbag格式原始数据。我们采用背景提取、离群点去除、地面投影和欧几里得聚类作为分割步骤，使用ESF和VFH描述子与随机森林、SVM和kNN分类器进行分类，并设计了基于加权速度的跟踪算法。在分类实验中，随机森林+ESF表现最优，且训练于简单序列的模型在复杂序列上泛化更好；跟踪实验中，MOTA值随训练/测试组合变化，MOTP在94mm至179mm之间，表明定位精度较高。数据集还包含工具用于基准评测。未来计划在更多室内大型场所中扩展序列。核心机制包括速度加权更新公式和MOTA/MOTP评估指标。\n\n### 关键要点\n- InLiDa数据集包含六个序列，采集于两个大型室内环境（走廊和大厅），场景复杂度递增。\n- 数据使用Velodyne VLP-16激光雷达获取3D点云，存储为PCD_V7格式，便于与PCL库协同。\n- 数据提供了点级标注，包含人和机器人的质心及包围立方体，并分配唯一ID以便跨帧跟踪。\n- 机器人（ID=0）作为移动代理加入场景，增加了遮挡和交互复杂性，模拟人机交互场景。\n- 每个序列附带20秒无动态对象的子序列，用于背景提取，原始数据也以rosbag格式提供。\n- 检测分割流程包含背景去除、离群点清除、地面投影和欧几里得聚类四个步骤。\n- 分类部分使用ESF和VFH全局描述子，并测试了随机森林、SVM（线性与Chi2核）和kNN（k=7）三种方法。\n- 实验表明随机森林+ESF组合精度最高，且使用简单序列（1-3）训练可以更好地泛化到复杂序列（5-6）。\n- 跟踪算法基于历史速度的加权平均，公式V=sum(v_i*w_i)更新速度，预测下一帧位置，并设置匹配阈值0.5m。\n- 跟踪性能采用MOTA和MOTP评估，MOTP在94-179mm之间，误差小于0.2m，表明定位相当精确。\n- 数据集和工具在<http://simdresearch.com/datasets/inlida>公开，用于人员检测与跟踪的基准评估。"
}