顾及系统误差的坐标转换方法研究

摘要

As Bursa model does not consider the local deformation and accumulation error in the coordinate system conversion , by ana-lyzing on the coordinate conversion errors , this paper applied the semi -parametric model to achieve measured coordinates in a region to transform by considering systematic errors as semi -parametric signals , estimating the nonparametric component of the check points, and comparison with Bursa model was carried out finally .The results show that the semi -parametric model could effectively eliminate systematic errors .And also to explore the effects of different methods of determining smoothing parameter on the results , the results show that under the same circumstances of regular matrix , different methods of determining smoothing parameter , the accuracy of coordinate transformation is different , but is improved with the models than with Bursa model .%针对Bursa模型在坐标系转换时没有顾及局部变形和累积误差的问题，通过对坐标转换误差进行分析，本文提出将由此产生的系统误差看作非参数信号的半参数估计，采用半参数模型对某一区域坐标进行解算，并对检核点非参数分量进行推估，与Bursa模型进行比较，结果表明半参数模型能够有效地消除系统误差。并探讨了不同确定平滑因子α的方法对坐标转换精度影响，计算结果表明，在正则矩阵R相同情况下，不同平滑因子确定方法得到的坐标转换精度有所不同，但均优于Bursa模型转换精度。