中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊

重庆交通大学学报(自然科学版) ›› 2018, Vol. 37 ›› Issue (11): 64-68.DOI: 10.3969/j.issn.1674-0696.2018.11.11

• 港口航道·水利水电·资源环境 • 上一篇    下一篇

BP神经网络在水下地形高程拟合的应用

彭中波, 高阳   

  1. (重庆交通大学 航运与船舶工程学院,重庆 400041)
  • 收稿日期:2017-12-05 修回日期:2018-09-17 出版日期:2018-11-19 发布日期:2018-11-19
  • 作者简介:彭中波(1975—),男,湖南长沙人,教授,博士,主要从事船舶装备监控与故障诊断技术方面的研究。E-mail:yumibobo114@hotmail.com。 通信作者:高阳(1993—),男,重庆人,硕士研究生,主要从事船舶装备监控与故障诊断技术方面的研究。E-mail:924179397@qq.com。

Application of BP Neural Network in Elevation Fitting of Underwater Terrain

PENG Zhongbo, GAO Yang   

  1. (College of Shipping and Naval Engineering, Chongqing Jiaotong University, Chongqing 401320, P.R.China)
  • Received:2017-12-05 Revised:2018-09-17 Online:2018-11-19 Published:2018-11-19

摘要: 运用MATLAB软件的神经网络模块建立BP神经网络,以工程河段实测水下地形图中测量点的X、Y坐标值作为输入层神经元,相对应的高程值作为输出层神经元。针对网络训练样本数据较多的特点,对多种训练函数的优劣进行比较,选取适合的训练函数,并进行大量训练实验,不断修正性能参数,并利用水下地形图测量点样本数据进行检验。实验结果表明:该模型对水下地形高程的高精度预测作用满足地形测量工作的工程要求,在实际工程测量中具有很好的应用价值。

关键词: 航道工程, BP神经网络, 高程, 拟合

Abstract: A BP neural network was established with neural network module of MATLAB software, and the value of X and Y coordinates of the measured points in the underwater topographic map of the engineering river reach were taken as the input layer neurons, and the corresponding elevation value was the output layer neuron. According to the mass network training sample data,the advantages of various training functions were compared to select the training function suitable for this study, and a large number of training experiments were carried out. The experimental results show that the model can predict the elevation of underwater topography with high accuracy and meet the engineering requirements of topographic survey. So it has a good application value in practical engineering measurement.

Key words: waterway engineering, BP neural network, elevation, fitting

中图分类号: