| 利用Kmeans-RBF神经网络提高重力格网化数据精度 |
| |
| 引用本文： | 黄子炎,王庆宾,李锁柱,李亮,冯进凯.利用Kmeans-RBF神经网络提高重力格网化数据精度[J].海洋测绘,2021(4):43-47. |
| |
| 作者姓名： | 黄子炎 王庆宾 李锁柱 李亮 冯进凯 |
| |
| 作者单位： | 战略支援部队信息工程大学 地理空间信息学院,河南 郑州 450001 ；61243 部队,新疆 乌鲁木齐 830002 |
| |
| 基金项目： | 国家自然科学基金(41774018)；地理信息工程国家重点实验室开放基金 (SKLGIE2019-M-1-3)；信息工程大学科研团队发展基金(f5206) |
| |
| 摘 要： | 针对重力测量数据在格网化过程中精度会被降低的问题,顾及空间重力异常和地形的强相关性,提出了三维Kmeans-RBF神经网络方法,该方法利用神经网络的复杂非线性映射学习能力进行推估建模,并在模型训练和推估时加入地形数据作为物理控制。最后基于美国爱达荷州地区的实测重力数据进行验证,实验结果表明:该方法相对于二维Kmeans-RBF神经网络方法和直接进行拟合推估的Kriging方法,实验区内精度分别提高了24.85%和44.84%。
|
| 关 键 词： | 重力异常 格网化 径向基函数神经网络 Kmeans聚类 深度学习 |
Using Kmeans-RBF neural network to improve the accuracy of grid gravity data |
| |
| HUANG Ziyan,WANG Qingbin,LI Suozhu,LI Liang,FENG Jinkai.Using Kmeans-RBF neural network to improve the accuracy of grid gravity data[J].Hydrographic Surveying and Charting,2021(4):43-47. |
| |
| Authors: | HUANG Ziyan WANG Qingbin LI Suozhu LI Liang FENG Jinkai |
| |
| Abstract: | Aiming at the problem that the accuracy of gravity data will be reduced in the process of gridding,considering the strong correlation between spatial gravity anomaly and terrain,this paper proposes a three-dimensional kmeans RBF neural network method,which uses the ability of neural network to learn complex nonlinear mapping for modeling,and adds terrain data as physical control in the process of training model and estimation.Finally,the measured gravity data in Idaho were used to verify the experimental results.Compared with the two-dimensional KMeans RBF neural network method and the Kriging method,the accuracy of the proposed method in the experimental area is improved by 24.85% and 44.84%,respectively. |
| |
| Keywords: | |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《海洋测绘》浏览原始摘要信息 |
| 点击此处可从《海洋测绘》下载免费的PDF全文 |
|
