内河渡船异常行为识别

目前,内河渡运监管模式主要是基于VTS、AIS的人力分析判断，属于被动监管，其效率难以应对日益严峻的水上安全监管形势.本文利用核密度估计对渡船历史AIS数据进行统计分析，挖掘渡船运动模式，得到渡船位置、航向和航速等运动特征的概率密度空间分布.在此基础上，建立基于位置异常和区域速度异常的渡船异常行为检测算法，并选择真实渡船AIS数据对算法进行检验.结果表明,该算法能够准确地辨识出渡船的异常行为，对水上监管具有辅助作用.
At present, the supervision mode of inland ferrying was conducted passively by mainly human analysis and judgement based on VTS and AIS, which is not favorable to reply increasingly serious situation of marine security supervision. The historical AIS data of the ferryboat was analyzed to explore its motion pattern and obtain the probability density spatial distribution of motion characteristics of location, course and speed by using the kernel density estimation. Then on this basis, the detection algorithm of the ferryboat abnormal behaviors was established for location and speed abnormity, and actual AIS data was used to verify the algorithm. Experimental results show that the proposed algorithm can accurately recognize the abnormal behaviors of the ferryboat, which is helpful to the marine supervision.