江西科学
江西科學
강서과학
JIANGXI SCIENCE
2012年
3期
308-310,415
,共4页
神经网络%GPS%大地高%正常高%精度
神經網絡%GPS%大地高%正常高%精度
신경망락%GPS%대지고%정상고%정도
: Artificial neural network%GPS height%Normal height%Precision
探讨了用神经网络方法转换GPS大地高为正高或正常高,给出了一种改进的BP神经网络拓扑结构和算法,并用GPS实测资料数据进行了计算分析,估算精度达到厘米级,且精度比较稳定,已知样本点的数量要求较少。
探討瞭用神經網絡方法轉換GPS大地高為正高或正常高,給齣瞭一種改進的BP神經網絡拓撲結構和算法,併用GPS實測資料數據進行瞭計算分析,估算精度達到釐米級,且精度比較穩定,已知樣本點的數量要求較少。
탐토료용신경망락방법전환GPS대지고위정고혹정상고,급출료일충개진적BP신경망락탁복결구화산법,병용GPS실측자료수거진행료계산분석,고산정도체도전미급,차정도비교은정,이지양본점적수량요구교소。
In this paper, A conversion of GPS height to normal height by means of neural network is proposed. The topological structure and the algorithm of the BP neural network are presented. Calcu- lated and analyzed by the measured data of GPS, the estimation precision to centimeter, and the pre- cision is relatively stable, the number of the known sample points requiring less.
