设为首页 | 加入收藏
黄奇华,吴雄斌,岳显昌,张兰.基于BP神经网络的高频地波雷达海流空间插值[J].海洋学报,2019,41(5):138-145
基于BP神经网络的高频地波雷达海流空间插值
Spatial interpolation of current mapped by HF surface wave radar using BP neural network
投稿时间:2018-06-22  
DOI:10.3969/ji.ssn.0253-4193.2019.05.013
中文关键词:  高频地波雷达  海流  空间插值  BP神经网络
英文关键词:high frequency ground wave radar  ocean surface current  spatial interpolation  BP neutral network
基金项目:国家自然科学基金项目(61771352);"十二五"国家高技术研究发展计划(863计划)资助项目(2012AA091701,2012AA091702)。
作者单位E-mail
黄奇华 武汉大学 电子信息学院, 湖北 武汉 430072  
吴雄斌 武汉大学 电子信息学院, 湖北 武汉 430072 xbwu@whu.edu.cn 
岳显昌 武汉大学 电子信息学院, 湖北 武汉 430072  
张兰 武汉大学 电子信息学院, 湖北 武汉 430072  
摘要点击次数: 987
全文下载次数: 744
中文摘要:
      高频地波雷达是海洋环境监测的重要手段,当前已经实现对海流的业务化观测,但是外部因素常引起海流空间探测的不连续性。为解决此问题,尽量保障区域数据的完整性和准确性,本文将BP神经网络技术与空间插值相结合,建立了海流的BP神经网络插值模型,并进行了针对实测数据的缺失插值仿真,通过与反距离权重法和线性插值法插值结果的对比,分析该模型在区域海流大面积缺失、流速整体较大和流速整体较小3个方面的性能。结果表明,BP神经网络插值模型的海流预测效果明显优于其他两种方法,且在流场数据大范围缺失下也取得了良好的效果。
英文摘要:
      High frequency (HF) ground wave radar is an important means of sea monitoring, HF radar routine observations of sea current have been operated for decades of years. Gaps in current data often occur due to external interferences. In order to ensure the integrity and accuracy of regional data, a back propagation (BP) neutral network interpolation model for ocean current is established by combining BP network technology and spatial interpolation. Two other interpolation methods, inverse distance weighted method and linear interpolation method, are adopted for comparison to validate the performance of the BP neutral network interpolation model. Simulations are conducted to analysis the performance of this new model in the cases of large areas of ocean current loss, large current velocity and relatively small current velocity. The results show that the prediction effect of the BP neutral network method is obviously better than the other two methods, and the new model also achieves good results in the absence of a wide range of current.
查看全文   查看/发表评论  下载PDF阅读器
关闭
微信公共账号