基于Multi-DOTS算法的船舶轨迹数据压缩研究

Compression of ship track data with Multi-DOTS algorithm

为在船舶轨迹数据的压缩过程中保留船舶运动特性,提出一种基于有向无环图的在线轨迹数据多步压缩(Multi-step Directed Acyclic Graph Based Online Trajectory Simplification, Multi-DOTS)算法。Multi-DOTS算法先对输入的船舶轨迹进行检测并划分,然后采用基于有向无环图的在线轨迹数据压缩(Directed Acyclic Graph Based Online Trajectory Simplification, DOTS)算法初步确定最优的局部子轨迹压缩路径,再以对应子轨迹类型的滑动统计窗口检测并保留有关键运动信息的船舶轨迹点,最终综合各局部压缩子轨迹得到全局压缩轨迹。以厦门港及附近水域船舶轨迹数据验证了算法的有效性,检验结果表明:当压缩率小于80%时,Multi-DOTS算法的动态误差约为道格拉斯-普克(Douglas-Peucker, DP)算法和DOTS算法的65.17%;在相同参数条件下对不同类型船舶的轨迹数据有稳定的压缩率。可见Multi-DOTS算法避免了参数的频繁切换,压缩后的船舶轨迹数据能更好地反映船舶运动特性,能满足船舶交通流特性分析及船舶行为模式识别等应用研究的需要。

An online trajectory processing algorithm based on multi-step directed acyclic graph(Multi-DOTS)algorithm is introduced for compressing ship track data.The Multi-DOTS algorithm acquires ship track data and segments the track according to the shape of the track and works out the paths for compressing each section data in the sense of locally optimization.The sections of the ship track are processed with sliding window corresponding to each track section and the track points with ship motion information is selected.The section data are compressed individually and then put together to form the compressed whole ship track.The algorithm is used to process track data of the ships at Xiamen port and adjacent waterway for verification.Experiments show that,when compression rate is lower than 80%,the dynamic error is 65.17%of the error with Douglas Peucker algorithm(DP)or pure DOTS algorithm.