基于深度卷积神经网络的船舶轨迹分类研究  

Research on Ship Trajectory Classification Based on Deep Convolution Neural Network

作　　者：常吉亮 谢磊[1] 魏志威 杨洋[1,2] 赵建伟 CHANG Jiliang;XIE Lei;WEI Zhiwei;YANG Yang;ZHAO Jianwei(National Engineering Research Center for Water Transport Safety, Wuhan University of Technology, Wuhan 430063, China;School of Naval Architecture, Ocean and Energy Power Engineering, Wuhan University of Technology, Wuhan 430063, China;China Institute of Marine Technology and Economy, Beijing 100081, China)

机构地区：[1]武汉理工大学国家水运安全工程技术研究中心,武汉430063 [2]武汉理工大学船海与能源动力工程学院,武汉430063 [3]中国船舶工业综合技术经济研究院,北京100081

出　　处：《武汉理工大学学报（交通科学与工程版）》2022年第1期160-165,共6页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家重点研发计划课题(2017YFC0804900,2017YFC0804904)。

摘　　要：为了提取船舶轨迹数据的空间特征,解决船舶轨迹分类及船舶轨迹所属航道识别问题,提出了一种基于深度卷积神经网络的船舶轨迹分类方法.考虑到经纬度数据难以准确描述船舶轨迹空间特征,将船舶轨迹数据转换为船舶轨迹图像数据,并建立数据集.构建基于深度卷积神经网络的船舶轨迹分类模型,使用人工标注的数据集开展训练.选取以经纬度数据为输入的全连接神经网络模型和SVM模型进行对比分析.结果表明:基于深度卷积神经网络船舶轨迹分类模型可以有效地区分不同航道内的船舶轨迹,所提方法是一种有效的船舶轨迹分类方法.In order to extract the spatial features of ship trajectory data and solve the problems of ship trajectory classification and channel identification,a ship trajectory classification method based on deep convolution neural network was proposed.Considering that latitude and longitude data is difficult to accurately describe the spatial characteristics of ship trajectory,the ship trajectory data was converted into ship trajectory image data,and a data set was established.A ship trajectory classification model based on deep convolution neural network was constructed,and the manually labeled data set was used for training.The fully connected neural network model and SVM model with latitude and longitude data as input were selected as control,and a comparative analysis was made.The results show that the ship trajectory classification model based on deep convolution neural network can effectively distinguish the ship trajectories in different channels,and the proposed method is an effective ship trajectory classification method.

关 键 词：智能交通 AIS 船舶轨迹分类 深度卷积神经网络 ResNet50 

分 类 号：U697.1[交通运输工程—港口、海岸及近海工程]