当前位置: X-MOL 学术Mobile Netw. Appl. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Adaptive Extraction and Refinement of Marine Lanes from Crowdsourced Trajectory Data
Mobile Networks and Applications ( IF 3.8 ) Pub Date : 2020-02-12 , DOI: 10.1007/s11036-019-01454-w
Guiling Wang , Jinlong Meng , Zhuoran Li , Marc Hesenius , Weilong Ding , Yanbo Han , Volker Gruhn

Crowdsourced trajectory data of ships provide the opportunity for extracting marine lane information. However, extracting useful knowledge from massive amounts of trajectory data is a challenging problem. Trajectory data collected from crowdsourcing can be extremely diverse in different areas and its quality might be very low. Moreover, the density distribution of the crowdsourced trajectory points is quite uneven in different areas. Furthermore, it is necessary to extract marine lanes with high extraction precision in offshore and nearshore water areas, but extraction precision can be lower in the open sea. We propose an adaptive approach for marine lane extraction and refinement based on grid merging and filtering to meet the challenges. In this paper, after pre-processing and clustering the trajectory data based on the density value of grids with a parallel GeoHash encoding algorithm, we propose a parallel grid merging and filtering algorithm based on a QuadTree data structure. The algorithm performs grid merging on the simplified grid data according to the density value of grid, then filters the merged grid data based on a local sliding window mechanism to get the marine lane grid data. Applying the Delaunay Triangulation on the marine lane grid data, the marine lane boundary information can be extracted with adaptive extraction precision. Experimental results show that the proposed approach can extract marine lanes with high extraction precision in offshore and nearshore water area and low extraction precision in open sea area.

中文翻译:

从众包轨迹数据自适应提取和优化海洋航道

船舶的众包轨迹数据为提取海上航道信息提供了机会。然而,从大量的轨迹数据中提取有用的知识是一个具有挑战性的问题。从众包收集的轨迹数据在不同区域可能非常不同,其质量可能很低。此外,众包轨迹点的密度分布在不同区域非常不均匀。此外,有必要在近海和近岸水域中提取具有高提取精度的海上航道,但是在公海中提取精度可能较低。我们提出了一种基于网格合并和过滤的自适应航道提取和优化方法,以应对挑战。在本文中,在使用网格GeoHash编码算法对基于网格密度值的轨迹数据进行预处理和聚类之后,提出了一种基于QuadTree数据结构的并行网格合并和过滤算法。该算法根据网格的密度值对简化后的网格数据进行网格合并,然后基于局部滑动窗口机制对合并后的网格数据进行过滤,得到航道网格数据。将Delaunay三角剖分应用于海道网格数据,可以以自适应提取精度提取海道边界信息。实验结果表明,该方法在海上和近岸水域中提取精度较高,而在公海中提取精度较低。我们提出了一种基于QuadTree数据结构的并行网格合并和过滤算法。该算法根据网格的密度值对简化后的网格数据进行网格合并,然后基于局部滑动窗口机制对合并后的网格数据进行过滤,得到航道网格数据。将Delaunay三角剖分应用于海道网格数据,可以以自适应提取精度提取海道边界信息。实验结果表明,该方法在海上和近岸水域中提取精度较高,而在公海中提取精度较低。我们提出了一种基于QuadTree数据结构的并行网格合并和过滤算法。该算法根据网格的密度值对简化后的网格数据进行网格合并,然后基于局部滑动窗口机制对合并后的网格数据进行过滤,得到航道网格数据。将Delaunay三角剖分应用于海道网格数据,可以以自适应提取精度提取海道边界信息。实验结果表明,该方法在海上和近岸水域中提取精度较高,而在公海中提取精度较低。然后基于局部滑动窗口机制对合并的网格数据进行过滤,以获取航道网格数据。将Delaunay三角剖分应用于海道网格数据,可以以自适应提取精度提取海道边界信息。实验结果表明,该方法在海上和近岸水域中提取精度较高,而在公海中提取精度较低。然后基于局部滑动窗口机制对合并的网格数据进行过滤,以获取航道网格数据。将Delaunay三角剖分应用于海道网格数据,可以以自适应提取精度提取海道边界信息。实验结果表明,该方法在海上和近岸水域中提取精度较高,而在公海中提取精度较低。
更新日期:2020-02-12
down
wechat
bug