基于BP神经网络的不完全测量数据处理方法研究  

The Approach of Incomplete Surveying Data Based On BP Neural Net

作　　者：任云志[1] 贺跃光[1] 吴弘[1] 姬方[1] 戴潇蕾[1] 

机构地区：[1]长沙理工大学交通运输工程学院,湖南长沙410114

出　　处：《现代测绘》2013年第1期9-11,15,共4页Modern Surveying and Mapping

摘　　要：由于受施工干扰或测量条件影响,致使测量数据不完全或测量数据缺失,增加数据分析难度,造成分析结果偏差。BP神经网络具有很好的非线性映射能力、自学习能力、自组织能力和自适应能力。为此,提出了运用BP神经网络算法对不完全测量数据进行处理的方法,得到正确平差结果。结果表明,BP神经网络算法无需知道系统模型,对待估计参数的先验信息要求较少,其估计结果较好地反映地表沉降规律。The construction disturbance or accident result the incomplete surveying data or missing surveying data, it will increase the difficulty of data analysis and the deviation of the analysis results. BP neural network has good nonlinear mapping ability, learning ability, self-organization ability and self-adaptive ability. In this paper, we propose a new data processing method which use BP neural network algorithm for incomplete surveying data to get the correct adjustment results to improve measurement accuracy. The results show that the BP neural network algorithm does not need to know the systematic model and requires less priori information of the estimated parameters. The estimation results of it can reflect the ground subsidence law better.

关 键 词：BP神经网络不完全测量数据数据处理地表沉降 

分 类 号：TU196[建筑科学—建筑理论]