从假设条件、估计准则和估值特性等方面比较最小二乘配置和普通Kriging法的异同。结合算例比较了配置法和普通Kriging法在不同观测噪声和数据分辨率水平下的推估精度及特性,算例1表明:相同数据分辨率下,观测噪声对普通Kriging法的估值影响大于配置法;相同观测噪声下,数据分辨率对配置法的估值影响大于普通Kriging法。算例2表明:在局部高程异常的拟合中,普通Kriging法也是一种有效推估方法。
The methods of least square collocation and ordinary Kriging are introduced, and then the similarities and differences between them in the supposing condition, estimate criterion and characters of estimate value are pointed out. The performance and accuracy estimated by collocation and ordinary Kriging are compared through an example of simulated gravity anomaly data with different observation errors and data resolution. The results show that the effect of observation error on the accuracy estimated by Kriging is greater than that by collocation under the same data resolution, and the effect of data resolution on the accuracy estimated by collocation is greater than that by Kriging under the same observation error. Finally, the two prediction techniques are applied to fit the height anomaly in a local area, the result indicates that the Ordinary Kriging is also an effective method for refining the local geoid.