{
  "paper_id": "de3f0207dd2727e29b37691f67ed10bc",
  "paper_title": "Deep Person Detection in 2D Range Data",
  "paper_authors": [
    "Lucas Beyer",
    "Alexander Hermans",
    "Timm Linder",
    "Kai O. Arras",
    "Bastian Leibe"
  ],
  "publication_date": "",
  "publication_venue": "ArXiv",
  "abstract": "Detecting humans is a key skill for mobile robots and intelligent vehicles in a large variety of applications. While the problem is well studied for certain sensory modalities such as image data, few works exist that address this detection task using 2D range data. However, a widespread sensory setup for many mobile robots in service and domestic applications contains a horizontally mounted 2D laser scanner. Detecting people from 2D range data is challenging due to the speed and dynamics of human leg motion and the high levels of occlusion and self-occlusion particularly in crowds of people. While previous approaches mostly relied on handcrafted features, we recently developed the deep learning based wheelchair and walker detector DROW. In this paper, we show the generalization to people, including small modifications that significantly boost DROW's performance. Additionally, by providing a small, fully online temporal window in our network, we further boost our score. We extend the DROW dataset with person annotations, making this the largest dataset of person annotations in 2D range data, recorded during several days in a real-world environment with high diversity. Extensive experiments with three current baseline methods indicate it is a challenging dataset, on which our improved DROW detector beats the current state-of-the-art.",
  "keywords": [
    "2D range data",
    "person detection",
    "deep learning",
    "DROW",
    "temporal integration",
    "mobile robots"
  ],
  "paper_institutions": [
    "RWTH Aachen University",
    "University of Bonn"
  ],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:46:12.750436Z",
  "template_tag": "simple",
  "summary": "### 摘要\n本文针对移动机器人和智能车辆在2D激光扫描数据中检测行人的问题，提出了一种基于深度学习的改进方法。传统的2D激光行人检测多依赖手工特征或基于腿部跟踪的方法，存在难以处理遮挡、动态变化和不同衣着等局限。作者在原有的DROW（基于深度学习的轮椅和助行器检测器）基础上，扩展了行人检测类别，并改进了投票机制和后处理步骤，同时通过引入全在线的多帧时间窗口融合方法，显著提升了检测性能。为了评估，作者扩展了DROW数据集，增加了大量行人标注，使其成为目前最大的真实场景2D激光行人检测数据集（超过10小时记录）。实验对比了三种现有基线方法，结果表明改进后的DROW（DROW3x）在性能上超越了当前最先进的方法，达到了较高的AUC值。该研究验证了从数据中直接学习而非手工设计特征的优越性，并展示了时间信息和里程计信息对检测的积极作用。\n\n### 关键要点\n- 本文首次将深度学习检测器DROW扩展到行人检测，并同时检测轮椅和助行器，实现了多类检测。\n- 改进的投票机制（联合投票）解决了原方法中检测位置离散化和缺乏置信度的问题，并提升了检测精度。\n- 通过增大网络规模（DROW3x）和采用更优的初始化、批归一化等技术，进一步提升了性能。\n- 提出了三种时间窗口生成方式（朴素、固定位置、加里程计修正），并验证了固定位置和里程计修正的有效性。\n- 时间融合策略中，晚期融合（LF）优于早期融合（EF），在加里程计修正后达到最佳效果。\n- 新数据集包含113个序列、464,013次扫描，其中24,012个标注扫描，有28,984个行人标注，是当前最大的真实场景2D激光行人数据集。\n- 与三个基线（Arras段分类、ROS腿部检测、Leigh联合腿部跟踪）对比，DROW3x在AUC上领先最强基线约10.9%。\n- 腿部跟踪方法在距离超过7米时性能急剧下降，而DROW和段分类方法可检测至15米。\n- 在定位精度评估（0.3米半径）下，腿部跟踪方法明显受影响，DROW和段分类方法几乎不受影响，显示DROW定位更精确。\n- 利用更大的训练集（合并训练和验证集）可进一步提升性能，说明模型具有更多可学习容量。\n- 通过Hyperopt自动调参，但发现自动调参可能过拟合验证集，手动合理参数有时表现更好，但DROW仍表现出鲁棒性。\n- 该研究强调强检测器对多目标跟踪系统的重要性，DROW是唯一能同时检测多类物体的强检测器。"
}