摘要
为更好地分析IGS连续运行参考站高程数据的变化规律及其变化趋势,以及预测将来一段时间内高程数据的变化,基于RBF正则化神经网络及小波滤波神经网络理论,以MATLAB7.0为平台对北京IGS站的高程分量数据进行GRNN函数逼近和小波分解逼近。通过对历年高程数据进行拟合和分层滤波,分析发现高程时间序列存在季节变化,其是以半周年项和年周项的季节性变化,其中年周期项比较明显。
In order to analyze the change law and tendency of height data of IGS continuous operation reference station and predict the future changes of the height data over a period of time, based on the theories of RBF regularization neural network and wavelet filter neural networks, the height component data of Beijing IGS station is dealt with the GRNN function approximation and wavelet decomposition approximation on the basis of the MATLAB7.0 platform. According to the fitting historical height data and hierarchical filtering, the seasonal changes were found in height time series, and seasonal changes included half year cycle changes and year cycle changes, and the year cycle was obvious.
出处
《大地测量与地球动力学》
CSCD
北大核心
2013年第5期136-139,共4页
Journal of Geodesy and Geodynamics
基金
国家基础测绘项目(50474032)
辽宁工程技术大学优秀青年基金(09-259)
