摘要
为提高船舶航行环境感知信息分类速度,研究基于大数据分析技术的船舶航行环境感知信息实时分类方法。采用工具层中的激光雷达传感器、风速/风向传感器、温湿度传感器,感知的船舶航行环境中障碍物、风速与风向、温湿度信息,在处理层中的Hadoop分布式大数据计算引擎中,由MapReduce并行大数据计算技术,将感知信息分块后,由map启动基于小波阈值的环境感知信息去噪方法,去除分块感知信息中噪声信息后,再启动K-最邻近分类器,计算去噪后分块感知信息样本与已知类型的感知信息隶属度,依据感知信息隶属度完成感知信息分类,最终通过reduce整合分类结果。经测试,本文方法对航行环境中感知信息分类时延仅2 s,延迟短,可实时分类船舶航行环境感知信息,且分类结果不存在信息混乱问题。
In order to improve the classification speed of ship navigation environment awareness information, a realtime classification method of ship navigation environment awareness information based on big data analysis technology is studied. Using the lidar sensor, wind speed/direction sensor, temperature and humidity sensor in the tool layer to perceive the obstacles, wind speed and direction, temperature and humidity information in the ship’s navigation environment, in the Hadoop distributed big data computing engine in the processing layer, MapReduce parallel big data computing technology is used to block the sensing information, and then the map starts the de-noising method of environment sensing information based on wavelet threshold, After removing the noise information in the block sensing information, start the K-nearest neighbor classifier, calculate the membership degree of the block sensing information samples and the known types of sensing information after noise removal, complete the classification of the sensing information according to the membership degree of the sensing information, and finally integrate the classification results through reduce. After testing, the proposed method has a short delay of only 2 s in classifying the perceptual information in the navigation environment, and can classify the perceptual information in the navigation environment in real time, and there is no information confusion in the classification results.
作者
汪洋
WANG Yang(Faculty of Computer and Information Technology,Wuhan Institute of Shipbuilding Technology,Wuhan 430050,China)
出处
《舰船科学技术》
北大核心
2022年第21期144-147,共4页
Ship Science and Technology
关键词
大数据分析
船舶航行
环境
感知信息
实时分类
去噪
big data analysis
ship navigation
environment
perceived information
real time classification
denoising
