基于小波去噪的灰色优化模型在变形监测中的应用研究  

Application of Gray Optimization Model Based on WaveletDenoising in Deformation Monitoring

作　　者：赵笠 ZHAO Li(Guangdong Heavy Industry Architectural Design Institute Co.,Ltd.,Guangzhou 510670,China)

机构地区：[1]广东省重工建筑设计院有限公司,广东广州510670

出　　处：《测绘与空间地理信息》2024年第5期221-224,共4页Geomatics & Spatial Information Technology

摘　　要：小波去噪在处理数据时能够剔除原始数据中的异常值,使得原始数据序列更加平滑。本文采用了基于小波阈值去噪和灰色模型的组合优化模型对建筑物沉降变形趋势进行预测研究,利用某工程项目中去噪后的前10期观测数据分别建立传统GM(1,1)模型、灰色Verhulst模型以及新陈代谢GM(1,1)模型,并对不同模型的预测效果进行综合分析。结果表明:基于小波阈值去噪的新陈代谢GM(1,1)模型预测精度最高,均方差为0.0449 mm,模型精度检验等级为Ⅰ等,因此,小波阈值去噪的新陈代谢GM(1,1)组合能够对建筑物沉降变形趋势进行更为科学准确的预测。Wavelet denoising can remove outliers in the original data when processing data,which makes the original data sequence smoother.In this paper,a combined optimization model based on wavelet threshold denoising and gray model is used to predict the trend of building settlement and deformation.The traditional GM(1,1)model,the gray Verhulst model and the metabolic GM(1,1)model were established using the first 10 periods of observation data after denoising in an engineering project,and the prediction effects of different models were comprehensively analyzed.The results show that the metabolic GM(1,1)model based on wavelet threshold denoising has the highest prediction accuracy,the mean square error is 0.0449 mm,and the model accuracy verification lev-el is I.Therefore,the metabolic GM(1,1)combination of wavelet threshold denoising can make more scientific and accurate predic-tions on the trend of building settlement and deformation.

关 键 词：小波阈值去噪 灰色VERHULST模型 新陈代谢GM(1 1)模型 精度检验 

分 类 号：P25[天文地球—测绘科学与技术] TB22[天文地球—大地测量学与测量工程]