基于深度学习的无人机河岸影像自动化识别研究  

Research on automated recognition of drone-based riverbank images based on deep learning

作　　者：杭震 陈瑶 HANG Zhen;CHEN Yao(Taizhou Port and Shipping Development Center,Taizhou 225300,China;Jiangsu Changtian Zhiyuan Transportation Technology Co.,Ltd.,Nanjing 210036,China)

机构地区：[1]泰州市港航事业发展中心,江苏泰州225300 [2]江苏长天智远交通科技有限公司,江苏南京210036

出　　处：《无线互联科技》2025年第4期22-27,共6页Wireless Internet Science and Technology

摘　　要：文章通过对内航河道相关沿线的无人机俯拍视频进行数据分析,结合交通执法部门维护航产航权需求,充分利用巡航无人机、神经网络等技术对航道进行智能化巡检。调查分析传统人工方式按月为周期对航段护岸、桥梁、锚地的巡查,存在实时性差、覆盖率低、费用昂贵等问题。针对上述问题,文章通过无人机巡航建立长纵深、多范围的内航河道智能识别巡检系统,解决港航养护业务中存在的重难点问题,实现航段监管、养护的自动化与智能化。This study analyzes aerial drone footage along inland waterways and integrates the requirements of traffic enforcement agencies for protecting navigation assets and rights.By fully utilizing patrol drones and neural network technologies,it develops an intelligent inspection system for navigation channels.Traditional manual inspection methods,conducted monthly,suffer from issues such as poor real-time performance,low coverage,and high costs when inspecting riverbanks,bridges and anchorages.To address these challenges,the study establishes a deep,wide-range intelligent recognition and inspection system for inland waterways through drone patrols.This system comprehensively tackles key and challenging issues in port and waterway maintenance,achieving full coverage of navigation sections above designated levels and enabling automation and intelligence in the monitoring and maintenance of navigation channels.

关 键 词：深度学习 智能感知巡航 无人机 河岸影像 神经网络 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]