城市勘测
城市勘測
성시감측
URBAN GEOTECHNICAL INVESTIGATION & SURVEYING
2014年
5期
15-18
,共4页
陈香萍%刘立龙%黎峻宇%蔡成辉%黄良珂
陳香萍%劉立龍%黎峻宇%蔡成輝%黃良珂
진향평%류립룡%려준우%채성휘%황량가
稳健估计%二次曲面拟合法%神经网络%高程异常%粗差
穩健估計%二次麯麵擬閤法%神經網絡%高程異常%粗差
은건고계%이차곡면의합법%신경망락%고정이상%조차
Robust Estimation%Two Curved Surface Fitting%Neural Network%Height Anomaly%Gross Error
为了研究稳健估计法在GPS高程拟合中的抗差能力,通过模拟在已知点中分别加入0.05m、0.1m、0.2m和0.3m的粗差以及在两个已知点中加入粗差,并采用二次曲面拟合法、BP神经网络模型和基于Huber法定权的稳健估计法进行对比。计算结果表明,稳健估计法应用于GPS高程拟合具有较好的抗差能力。
為瞭研究穩健估計法在GPS高程擬閤中的抗差能力,通過模擬在已知點中分彆加入0.05m、0.1m、0.2m和0.3m的粗差以及在兩箇已知點中加入粗差,併採用二次麯麵擬閤法、BP神經網絡模型和基于Huber法定權的穩健估計法進行對比。計算結果錶明,穩健估計法應用于GPS高程擬閤具有較好的抗差能力。
위료연구은건고계법재GPS고정의합중적항차능력,통과모의재이지점중분별가입0.05m、0.1m、0.2m화0.3m적조차이급재량개이지점중가입조차,병채용이차곡면의합법、BP신경망락모형화기우Huber법정권적은건고계법진행대비。계산결과표명,은건고계법응용우GPS고정의합구유교호적항차능력。
In order to study the error resistance capability of the robust estimation method in GPS height fitting, through the simulation of gross errors in 0.05m,0.1m,0.2m and 0.3m were added in the known points and in the two known points with error,then compare application of the two curved surface fitting,BP neural network model and robust estimation method based on Huber given the right.The results show that,the robust estimation method is applied to GPS elevation fitting has good robust capability.