{
  "paper_id": "ef5140a4c5c68df94e84962f156be052",
  "paper_title": "A Novel Approach of Counting People Based on Stereovision and DSP",
  "paper_authors": [
    "Zhu Qiuyu",
    "Tang Li",
    "Jiang Yiping",
    "Deng Wei-jun"
  ],
  "publication_date": "",
  "publication_venue": "2010 The 2nd International Conference on Computer and Automation Engineering (ICCAE)",
  "abstract": "A novel approach for counting people passing through a region where is surveilled by a parallel-fixed binocular camera is proposed in this paper. Firstly, two virtual counting lines are set to obtain four line space-time gray images. Then an integrated algorithm based on binocular stereovision and regional gray projection is proposed to extract and split overlapped moving objects in complex counting environment. At last, the object matching and counting is accomplished in space-time disparity maps. The counting algorithm takes full advantage of the objects' 2D and 3D information, and is implemented on the DSP hardware system. The practical test results show that the proposed system can operate in real-time and achieve a counting accuracy of over 96% in crowded people environment.",
  "keywords": [
    "people counter",
    "disparity map",
    "object detection",
    "object segmentation",
    "stereovision",
    "TMS320DM642"
  ],
  "paper_institutions": [
    "Shanghai University",
    "Shanghai Transportation Investment Information Technology CO., LTD"
  ],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:47:44.486052Z",
  "template_tag": "simple",
  "summary": "### 摘要\n\n本文提出了一种基于立体视觉和数字信号处理器（DSP）的人群计数新方法，旨在解决复杂监控环境下（如拥挤人群、光照变化和严重阴影）单目相机难以准确提取行人目标的问题。该方法采用平行固定的双目相机俯视监控区域，设置两条虚拟计数线（内线L1和外线L2），沿时间轴排列获取四幅线时空灰度图像。算法首先利用视差公式p=bf/d，通过基于归一化互相关（NCC）的1-D窗口匹配计算密集视差图，并根据人的高度设定阈值检测目标区域；然后针对拥挤场景中多人粘连问题，提出结合灰度投影的物体分割方法，根据目标宽度和面积分为三种情况分别处理；最后通过代价函数（综合平均视差、水平距离和时间距离）在内外线视差图中进行匹配，并根据出现顺序判断方向。系统基于TMS320DM642 DSP硬件实现，以25帧/秒实时运行。实验在两种环境（理想背景E1和复杂场景E2）和三种拥堵情况（单人连续、两人并排、多人同时通过）下进行，结果表明误差率均低于4%，计数准确率超过96%，验证了该方法对光照变化、阴影干扰和人群拥挤的鲁棒性。未来可进一步优化算法以适用于更广泛的应用场景和实时约束条件。核心创新在于融合2D灰度和3D深度信息，利用线时空图像降低计算量，并通过投影直方图有效分割重叠目标。\n\n### 关键要点\n\n- 系统采用平行固定的双目相机垂直向下安装，可显著减少行人重叠问题。\n- 设置两条虚拟计数线（L1和L2），沿时间轴排列最新100行数据生成线时空图像，分别来自左右相机共四幅。\n- 视差计算使用基于归一化互相关（NCC）的1-D窗口匹配，窗口沿极线方向，并通过二次插值获得亚像素精度。\n- 目标检测依据高度信息，通过设定视差阈值T_d分割头部区域，消除背景干扰。\n- 针对多人粘连，提出灰度投影与视差结合的分割方法：按宽度和面积分为三种情况，分别采用投影分割、仅视差分割或丢弃。\n- 灰度投影函数h(n)利用非零视差像素均值计算，以头发较暗的假设进行分割。\n- 匹配过程采用代价函数F(x,y_n)=aS+bD_x+cD_y，权重a=0.4、b=0.4、c=0.2，选择最小代价为最优匹配。\n- 方向判断依据目标在内外线视差图中出现的先后顺序。\n- 硬件采用TI TMS320DM642 DSP芯片，配合红外LED照明实现夜间低光环境下的高质量灰度图像采集。\n- 软件使用C语言和汇编语言在CCS及DSP/BIOS环境下开发，系统帧率可达25帧/秒。\n- 实验设置三种拥堵情况（单人连续、两人并排、多人同时）和两种环境（理想E1、复杂E2），各测试样本误差率均在-3.29%至+3.81%之间，总体误差率低于4%。\n- 计数准确率超过96%，系统对光照变化、阴影和拥挤人群具有良好鲁棒性，满足实时性和稳定性要求。"
}