GPS高程迭加拟合模型的研究

Research of combined model for GPS height

为提高GPS高程拟合模型在逼近高程异常曲面时的逼近精度和可信度,探讨了GPS高程转换的迭加模型(遗传神经网络模型和神经网络综合模型)。通过实测的GPS水准数据对迭加模型和单一模型进行分析比较,结果表明迭加模型逼近高程的精度和可靠性均高于单一模型。

In order to improve approximating precision and reliability of GPS leveling of imitated model in approxi-mating anormal height curved surface, the paper discussed the combined model of GPS elevation height to normal height, such as the genetic neural network model and neural network combined model. We compared the two combined models with single models. The results indicate that the precision and reliability are better than any single model.