改进BP神经网络算法用于GPS高程转换的应用

Application of Improved BP Neural-Network Algorithm to GPS Height Tansformation

介绍了用改进的BP神经网络算法转换GPS高程为正常高,并以青东煤矿首采工作面地表移动观测站的GPS数据为例,与标准的BP算法和多项式曲面拟合方法作比较,改进的BP算法在转换GPS高程中减少了BP神经网络的训练时间,提高了转换效率,且精度良好,可用于GPS高程转换。

Taking GPS data derived from observation station of surface movement in the first working face of Qingdong Coal Mine as an example,the conversion method of GPS height to normal height with the improved BP neural-network algorithm was investigated.The results show that the improved BP algorithm can reduce BP neural network training time and improve conversion efficiency with the higher precision in the procedure of GPS height conversion compared with the standard BP algorithm and the polynomial surface fitting.The improved BP algorithm can be used in the GPS height conservation.