无人机立体视觉识别船舶航行风险仿真
Simulation of Ship Navigation Risk Based on UAV Stereo Vision Recognition
刘小飞 1李明杰 1喻佳2
作者信息
- 1. 三亚学院信息与智能工程学院,海南 三亚 572000;三亚学院陈国良院士工作站,海南 三亚 572000
- 2. 华东交通大学信息工程学院,江西 南昌 330013
- 折叠
摘要
由于船舶交通受碰撞、搁浅、走锚等随机干扰因素的影响,导致对船舶航行目标的跟踪与风险预警具有较大难度.为增强海上航行的安全性,提出一种基于无人机立体视觉的船舶航行风险识别方法.引入双目视觉立体技术,获得目标船舶在航行过程中的图像数据.利用卡尔曼预测器优化连续性自适应均值漂移算法,跟踪目标船舶,通过高斯混合模型识别船舶航行风险.实验结果表明,研究方法对不同海域的船舶跟踪曲线与船舶实际航行曲线具有较高拟合度,且上述方法的船舶航行风险识别正确率高于 90%,且误报率低于0.2%,说明提出方法的应用可靠性较高.
Abstract
Because ship traffic is affected by random interference factors,it is difficult to track and warn the risk of ship navigation.In order to enhance the safety of maritime navigation,a method of identifying ship navigation risk is proposed based on UAV stereo vision.Firstly,binocular vision stereo technology was introduced to obtain the image data of the target ship during navigation.Then,a Kalman predictor was used to optimize the continuous adaptive mean shift algorithm and track the target ship.Finally,the navigation risk of the ship was identified by the Gaussian mixture model.Experimental results show that the proposed method has a high degree of fit between ship tracking curves in different sea areas and actual curves.Meanwhile,the accuracy rate of ship navigation risk identification is higher than 90%,and the false alarm rate is lower than 0.2%,proving the application reliability of the method.
关键词
无人机立体视觉/立体匹配/卡尔曼预测器/连续性自适应均值漂移算法/船舶航行风险识别Key words
UAV stereoscopic vision/Stereo matching/Kalman predictor/Continuous adaptive mean shift algo-rithm/Ship navigation risk identification引用本文复制引用
基金项目
海南省自然科学基金(622RC734)
海南省自然科学基金(621RC1077)
出版年
2024
NETL
NSTL
万方数据