The determination of vertical component plays a fundamental role in the initial phase of engineering applications. However, its acquisition is technically and economically demanding, mainly due to the precise heights ...The determination of vertical component plays a fundamental role in the initial phase of engineering applications. However, its acquisition is technically and economically demanding, mainly due to the precise heights relative to a reference surface, such as the mean sea level. The Cokriging technique is a necessary input for the calculation of the vertical component of the geodetic control points measured by GNSS, and it requires less auxiliary data and uses complementary available variables for the calculation.Therefore, the main goal is to use Cokriging to establish a geoid undulation prediction model for the rural area of the canton of Guayaquil, Ecuador. Ordinary, Residual and Universal Cokriging and Kriging techniques were used to compare their results and select the best for achieving accuracy. The validation of the application techniques yielded that Universal Cokriging was the most accurate with an RMSE of 8 cm and RSR of 2 cm, obtained just with 66 samples(20% of the dataset). Furthermore, considering the comparison with other regional geoid undulation models, the proposed model increased the accuracy of the results by a ratio of 9.68 and 6.96 in relation to EGM96 and EGM08, respectively.展开更多
文摘The determination of vertical component plays a fundamental role in the initial phase of engineering applications. However, its acquisition is technically and economically demanding, mainly due to the precise heights relative to a reference surface, such as the mean sea level. The Cokriging technique is a necessary input for the calculation of the vertical component of the geodetic control points measured by GNSS, and it requires less auxiliary data and uses complementary available variables for the calculation.Therefore, the main goal is to use Cokriging to establish a geoid undulation prediction model for the rural area of the canton of Guayaquil, Ecuador. Ordinary, Residual and Universal Cokriging and Kriging techniques were used to compare their results and select the best for achieving accuracy. The validation of the application techniques yielded that Universal Cokriging was the most accurate with an RMSE of 8 cm and RSR of 2 cm, obtained just with 66 samples(20% of the dataset). Furthermore, considering the comparison with other regional geoid undulation models, the proposed model increased the accuracy of the results by a ratio of 9.68 and 6.96 in relation to EGM96 and EGM08, respectively.
