一种适用于GNSS变形监测的滑动滤波模型  

Moving filtering model for GNSS deformation monitoring

作　　者：魏春晓 李博峰 WEI Chunxiao;LI Bofeng(College of Surveying and Geo-Informatics,Tongji University,Shanghai 200092,China)

机构地区：[1]同济大学测绘与地理信息学院

出　　处：《导航定位学报》2019年第3期125-130,共6页Journal of Navigation and Positioning

基　　金：国家自然科学基金项目(41874030,41622401,41574031);上海市科技委员会科技创新行动计划项目(17511109501)

摘　　要：针对GNSS变形监测数据处理中随机游走模型对局部时段快速变形带来的异常扰动抵抗能力较差,而常速度模型处理稳定变形的精度较低的问题,提出采用滑动滤波模型描述变形体的实际变形状态,并利用最小描述长度准则评估随机游走模型、常速度模型和滑动滤波模型在变形监测中的应用效果。结果表明,监测点稳定情况下滑动滤波模型平面精度可达到5 mm以内,且采用最小描述长度准则,能有效确定最优的滑动滤波模型,实现异常变形探测1 s内快速预警。In order to solve the problem that the random walk model has poor resistance to abnormal disturbance caused by rapid local deformation,while the constant velocity model has low accuracy in dealing with stable deformation,the moving filtering model is used to describe the actual deformation state of the deformation body.The minimum description length criterion is used to evaluate the random walk model,the constant velocity model and the moving filtering model in deformation monitoring.The results show that the plane accuracy of the moving filtering model can be less than 5 mm under the condition of stable deformation,and the optimal moving filtering model can be effectively determined by using the minimum description length criterion,to realize the rapid early warning of abnormal deformation detection within 1 s.

关 键 词：GNSS 变形监测 滑动滤波 最小描述长度 卡尔曼滤波 

分 类 号：P228[天文地球—大地测量学与测量工程]