考虑欧拉矢量的BP神经网络模型建立区域地壳运动速率场  

Building up Regional Crustal Movement Velocity Field with BP Neural Network Base on Euler Vector

作　　者：李冲[1,2] 李建成[1] 黄瑞金 谭理 

机构地区：[1]武汉大学测绘学院,湖北武汉430079 [2]四川省测绘产品质量监督检验站,四川成都610041

出　　处：《武汉大学学报（信息科学版）》2014年第3期362-366,共5页Geomatics and Information Science of Wuhan University

基　　金：国家自然科学基金资助项目(40637034;41210006;41274005)~~

摘　　要：讨论了常用的欧拉矢量模型和函数拟合模型的优缺点,提出了基于欧拉矢量的BP神经网络模型。该模型运用欧拉矢量的地学性质,结合BP神经网络在处理需要同时考虑许多因素和条件的、不确定和模糊的信息时的优势,可以较好地区分块体整体的刚性旋转及内部的弹性形变。经实例验证,取得较好的精度。In regional crustal movement research, mathematical model is always used to estimate with- out observation the points which need attention. Then, we can build up a relatively even and meaning- ful regional crustal movement velocity field. In this paper we analyze the strengths and weaknesses of the common Euler and function models, and propose a new model with BP neural network based on Euler vector. This proposed model uses the geological properties of a Euler vector and the superiority of BP neural network. It considers the various influences and uncertain information found in data pro- cessing. Therefore, the model can distinguish inner elastic strain from rigid-body of the plate. The proposed model obtains precision through specimen verification.

关 键 词：地壳运动 速率场 欧拉矢量 BP神经网络 

分 类 号：P227[天文地球—大地测量学与测量工程]