基于众包的室内地图自动构建方法研究与实现
发布时间：2019-01-28 21:25
【摘要】：基于位置的服务和应用,如基于位置的社交网络、游戏和广告等,在过去十多年里得到了快速增长。这得力于智能移动设备的普遍使用和定位技术的发展。这些基于位置的服务和应用通常使用地图来显示用户位置。户外位置服务提供商提供了几乎所有地区的室外街道地图,但室内平面图的发展目前仍非常有限,极大影响了基于室内定位的服务和应用的快速发展和部署。目前绝大多数室内应用都依赖于手动创建的室内平面图。手动添加、编辑和维护大量的建筑物平面图需要巨大的成本和努力。为解决上述问题,本文提出一种基于智能手机和众包方式的室内地图自动构建方法。该方法通过手机传感器采集人在室内行动的相关数据,通过行人航位推算算法计算得到人的步数、步长和方向,从而确定人在每一步的位置信息,并由此获得人的行走轨迹。为了准确高效构建地图,本文采用众包方式获取大量行人的轨迹信息,通过对大量轨迹数据的分析,构建出室内地图信息。为了检测室内标志点,本文提出一种有效的上下楼位置检测算法和房间门口位置检测算法。上下楼检测基于气压数据斜率变化和加速度量级。房间门口位置检测结合陀螺仪和WiFi进行,并通过基于密度的聚类算法进一步确定门口位置。根据门口位置,本文对大量的行人轨迹数据进行分段和聚类处理。对房间类型轨迹,本文使用?-shape算法构建出房间形状和大小。对走廊类型轨迹,本文采用主成分分析算法构建出走廊的长和宽。根据轨迹的位置坐标,将构建出的房间和走廊拼接成完整的室内地图,从而完成室内地图自动构建。本文对该方法进行了实现,并在实际环境中进行了实验。实验结果显示,在具有多房间的室内地图自动构建实验中,房间位置的平均误差为1.96m,走廊长宽的平均误差为1.85m。本方法构建的地图基本能反应真实地图中各个房间和走廊的相对位置关系,以及房间在走廊上的位置顺序。本文提出的基于众包的室内地图自动构建方法切实可行,实现了室内地图的自动构建。
[Abstract]:Location-based services and applications, such as location-based social networks, games and advertising, have grown rapidly over the past decade. This is due to the widespread use of intelligent mobile devices and the development of positioning technology. These location-based services and applications typically use maps to display user locations. Outdoor location service providers provide outdoor street maps in almost all areas, but the development of indoor plans is still very limited, which greatly affects the rapid development and deployment of indoor location-based services and applications. At present, most indoor applications rely on the manual creation of indoor plans. Manually adding, editing, and maintaining a large number of building plans requires enormous cost and effort. In order to solve the above problems, this paper presents an automatic building method of indoor map based on smart phone and crowdsourcing. In this method, the mobile phone sensor is used to collect the relative data of human movement in the room, and the pedestrian path calculation algorithm is used to calculate the step number, step size and direction of the person, so as to determine the position information of the person at each step, and thus obtain the human walking track. In order to construct the map accurately and efficiently, this paper uses crowdsourcing method to obtain a large number of pedestrian track information, through the analysis of a large number of track data, the indoor map information is constructed. In order to detect indoor markers, this paper presents an effective algorithm for detecting the position of the upper and lower floors and the location of the door of the room. The upper and lower building detects the magnitude of slope and acceleration based on air pressure data. The door position detection is carried out with gyroscope and WiFi, and the location of the door is further determined by density-based clustering algorithm. According to the location of the doorway, this paper deals with a large number of pedestrian trajectory data segmentation and clustering. This paper uses the?-shape algorithm to construct the shape and size of the room. In this paper, the length and width of corridor are constructed by principal component analysis (PCA) algorithm. According to the position coordinate of the track, the room and corridor are assembled into a complete indoor map, and the indoor map is constructed automatically. In this paper, the method is implemented, and the experiment is carried out in the actual environment. The experimental results show that the average error of room position is 1.96 m and the average error of corridor length and width is 1.85 m in the experiment of automatic building indoor map with multiple rooms. The map constructed by this method can basically reflect the relative position relationship of each room and corridor in the real map, as well as the location order of the room on the corridor. The automatic building method of indoor map based on crowdsourcing is feasible, and the automatic construction of indoor map is realized.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P283.7
本文编号：2417332
[Abstract]:Location-based services and applications, such as location-based social networks, games and advertising, have grown rapidly over the past decade. This is due to the widespread use of intelligent mobile devices and the development of positioning technology. These location-based services and applications typically use maps to display user locations. Outdoor location service providers provide outdoor street maps in almost all areas, but the development of indoor plans is still very limited, which greatly affects the rapid development and deployment of indoor location-based services and applications. At present, most indoor applications rely on the manual creation of indoor plans. Manually adding, editing, and maintaining a large number of building plans requires enormous cost and effort. In order to solve the above problems, this paper presents an automatic building method of indoor map based on smart phone and crowdsourcing. In this method, the mobile phone sensor is used to collect the relative data of human movement in the room, and the pedestrian path calculation algorithm is used to calculate the step number, step size and direction of the person, so as to determine the position information of the person at each step, and thus obtain the human walking track. In order to construct the map accurately and efficiently, this paper uses crowdsourcing method to obtain a large number of pedestrian track information, through the analysis of a large number of track data, the indoor map information is constructed. In order to detect indoor markers, this paper presents an effective algorithm for detecting the position of the upper and lower floors and the location of the door of the room. The upper and lower building detects the magnitude of slope and acceleration based on air pressure data. The door position detection is carried out with gyroscope and WiFi, and the location of the door is further determined by density-based clustering algorithm. According to the location of the doorway, this paper deals with a large number of pedestrian trajectory data segmentation and clustering. This paper uses the?-shape algorithm to construct the shape and size of the room. In this paper, the length and width of corridor are constructed by principal component analysis (PCA) algorithm. According to the position coordinate of the track, the room and corridor are assembled into a complete indoor map, and the indoor map is constructed automatically. In this paper, the method is implemented, and the experiment is carried out in the actual environment. The experimental results show that the average error of room position is 1.96 m and the average error of corridor length and width is 1.85 m in the experiment of automatic building indoor map with multiple rooms. The map constructed by this method can basically reflect the relative position relationship of each room and corridor in the real map, as well as the location order of the room on the corridor. The automatic building method of indoor map based on crowdsourcing is feasible, and the automatic construction of indoor map is realized.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P283.7
【参考文献】
中国期刊全文数据库 前5条
1 许华燕;李志辉;王金月;;室内地图应用现状及展望[J];测绘通报;2014年09期
2 徐飞;吕龙吟;秦子文;;基于LBS的无线定位技术方案研究[J];河南科技;2014年17期
3 张兰;王光霞;袁田;彭克曼;;室内地图研究初探[J];测绘与空间地理信息;2013年09期
4 刘兆宏;王科;丰江帆;夏英;;矢量室内地图建模与制作方法[J];数字通信;2012年04期
5 宋顺林，，詹永照，薛安荣，伏劲松;三维人体行走模型的研究与实现[J];计算机学报;1996年06期
本文编号：2417332
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2417332.html
