摘要自动客流计数(Automatic Passenger Counting，APC)是指无需人工干预，通过一定的检测方法，对视场内的人流信息进行自动统计的技术。自动客流计数也是智能交通系统(Intelligent Transportation System, ITS)的重要部分，实时采集到的乘客上下车人数、时间和地点的信息，可以为优化公交调度提供决策依据，从而提高公交系统的运营效率，节约商业运作成本。本文先简单介绍了目前应用比较广泛的客流计数方法，针对现有方法的不足，通过对Xtion体感器采集到的3D深度信息的分析，提出一种新的基于三维数据的客流计数方法 。该方法利用Xtion体感器传出的深度影像和彩色影像，结合Visual Studio软件和OpenCV、OpenNI开发包，由灰度级、类椭圆性、占空比等判决依据确定人物头部，建立头部运动航迹并进行跟踪计数，从而实现上下车客流的自动统计。同时，该方法可与GPS系统结合，实时跟踪监测客流信息。实验结果表明，该方法能够对视场内的人物头部进行快速有效的识别跟踪，具有较好的实时性和较高的计数精度,室内精度可达96%。64922

毕业论文关键词 自动客流计数 三维数据 目标识别 目标跟踪

毕业设计说明书（论文）外文摘要

Title   The design of automatic passenger counting system  based on 3D data                                                 

Abstract Automatic passenger counting(APC) technology can automatically count the number of the passengers by using some testing method. Automatic passenger counting is also one of the important part of the intelligent transportation system. The real-time collection of information about the number, time and position of passengers going up and down the buses,  which can provide the basis for optimizing bus scheduling and thus to achieving a more efficient public transport system and saving business operation cost. Firstly, the paper briefly introduces the prevalent automatic passenger counting methods, for the deficiencies of these methods and according to the analyses of the 3D depth information collected by the Xtion body sensor, we proposed a new passenger counting method based on the 3D data. This method takes advantage of the depth videos and color videos outputted by the Xtion body sensor, combines with Visual Studio, OpenCV and OpenNI software to find out the heads of people by gray level, concavity, convexity and duty ratio. Then we establish the movement tracks of the heads, track them and count its number so that we can get the number of passengers going up and down the buses. Simultaneously, real-time tracking and monitoring for passengers is probably if we combine this method with the GPS system. The experimental results show that the proposed method can detect and track heads of people effectively. At the same time, it meets the demand of real-time and high-accuracy application and the indoor accuracy up to 96%. 

Keywords  Automatic Passenger Counting   3D data   Object detection    Object tracking    

目   次

1  绪论··   1

    1.1  研究的背景与意义    1

  1.2  客流统计发展现状    1

  1.3  本文研究内容和方法·    3

2  硬件和软件介绍··   4

  2.1  Xtion体感器   4

  2.2  OpenCV和OpenNI··   7

  2.3  Visual Studio··  10

3  人物头部检测原理··· 10