面向移动对象的轨迹数据压缩算法研究
发布时间：2018-08-08 13:58
【摘要】：轨迹数据中蕴含着丰富的空间信息及时间信息,具有进一步分析、处理及利用的价值。随着定位技术的迅速发展和日渐普及,使得采集移动对象的轨迹数据变得容易。然而,随着轨迹采集终端的大量使用,原始轨迹数据量急剧增长,且冗余严重,给轨迹数据的存储、传输及进一步分析处理带来了极大的压力。因此,有效剔除原始轨迹中的冗余信息,实现轨迹数据压缩具有重要的意义。本文针对离线轨迹数据,综合考虑轨迹数据的速度、方向及位置特征,在保证原有轨迹重要信息不丢失的基础上,为实现各级别压缩率下对复杂轨迹数据的压缩,提出三种离线轨迹数据压缩算法:(1)基于速度特性的角度偏移量轨迹数据压缩算法。该算法根据轨迹数据的速度特征和方向特征,通过比较预设的速度阈值和角度阈值删除冗余,较好地保留了原始轨迹中速度变化较大及方向发生改变的轨迹点;(2)基于网格的轨迹数据压缩算法。该算法使用网格覆盖整条轨迹,通过计算各网格内最早轨迹点与其他轨迹点的时间差,剔除小于预设时间阈值的轨迹点,并利用弗里曼链码的编码方式进一步压缩相邻网格内方向信息冗余的轨迹点;(3)基于局部加权线性回归的轨迹数据压缩算法。该算法利用线性变化的曲线模拟非线性变化的轨迹,采用局部加权线性回归算法对原始轨迹进行曲线拟合,去除原始轨迹上偏离拟合曲线小于给定距离阈值的轨迹点。本文最后对三个算法分别进行了验证及误差比较与分析,结果表明,本文算法是可行的、有效的。
[Abstract]:The trace data contains abundant spatial information and time information, which has the value of further analysis, processing and utilization. With the rapid development and popularity of positioning technology, it is easy to collect track data of moving objects. However, with the extensive use of trajectory acquisition terminals, the original trajectory data volume increases rapidly, and the redundancy is serious, which brings great pressure to the storage, transmission and further analysis and processing of trajectory data. Therefore, it is of great significance to effectively eliminate redundant information from the original trajectory and achieve trajectory data compression. In this paper, we consider the velocity, direction and position characteristics of the off-line trajectory data synthetically, on the basis of ensuring that the important information of the original trajectory is not lost, in order to realize the compression of the complex trajectory data under the compression ratio of each level. Three off-line trajectory data compression algorithms are proposed: (1) angular offset trajectory data compression algorithm based on velocity characteristics. According to the velocity and direction characteristics of the trajectory data, the redundancy is removed by comparing the preset velocity threshold and the angle threshold. The trajectory points in the original trajectory are well preserved. (2) the grid-based trajectory data compression algorithm. The algorithm uses mesh to cover the whole track, and by calculating the time difference between the earliest trace points and other locus points in each grid, the trajectory points which are less than the preset time threshold are eliminated. The Freeman chain code is used to further compress the redundant locus points of direction information in adjacent grids. (3) the locus data compression algorithm based on local weighted linear regression. The algorithm uses the curve of linear variation to simulate the trajectory of nonlinear change, and uses the local weighted linear regression algorithm to fit the original trajectory, and removes the trajectory points on the original trajectory that deviate from the fitting curve less than the threshold of the given distance. Finally, the three algorithms are verified and the error is compared and analyzed. The results show that the algorithm is feasible and effective.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P208
[Abstract]:The trace data contains abundant spatial information and time information, which has the value of further analysis, processing and utilization. With the rapid development and popularity of positioning technology, it is easy to collect track data of moving objects. However, with the extensive use of trajectory acquisition terminals, the original trajectory data volume increases rapidly, and the redundancy is serious, which brings great pressure to the storage, transmission and further analysis and processing of trajectory data. Therefore, it is of great significance to effectively eliminate redundant information from the original trajectory and achieve trajectory data compression. In this paper, we consider the velocity, direction and position characteristics of the off-line trajectory data synthetically, on the basis of ensuring that the important information of the original trajectory is not lost, in order to realize the compression of the complex trajectory data under the compression ratio of each level. Three off-line trajectory data compression algorithms are proposed: (1) angular offset trajectory data compression algorithm based on velocity characteristics. According to the velocity and direction characteristics of the trajectory data, the redundancy is removed by comparing the preset velocity threshold and the angle threshold. The trajectory points in the original trajectory are well preserved. (2) the grid-based trajectory data compression algorithm. The algorithm uses mesh to cover the whole track, and by calculating the time difference between the earliest trace points and other locus points in each grid, the trajectory points which are less than the preset time threshold are eliminated. The Freeman chain code is used to further compress the redundant locus points of direction information in adjacent grids. (3) the locus data compression algorithm based on local weighted linear regression. The algorithm uses the curve of linear variation to simulate the trajectory of nonlinear change, and uses the local weighted linear regression algorithm to fit the original trajectory, and removes the trajectory points on the original trajectory that deviate from the fitting curve less than the threshold of the given distance. Finally, the three algorithms are verified and the error is compared and analyzed. The results show that the algorithm is feasible and effective.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P208
【参考文献】
相关期刊论文 前10条
1 程倩;丁云峰;;基于路网的GPS轨迹在线压缩方法[J];计算机系统应用;2016年06期
2 樊庆富;张磊;刘磊军;鲍苏宁;房晨;;基于偏移量计算的在线GPS轨迹数据压缩[J];计算机工程与应用;2017年08期
3 刘磊军;房晨;张磊;鲍苏宁;;基于运动状态改变的在线全球定位系统轨迹数据压缩[J];计算机应用;2016年01期
4 许佳捷;郑凯;池明e，
本文编号：2172002
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2172002.html
