| Summary: | 碩士 === 國立中興大學 === 土木工程學系所 === 98 === The elevation data is used to be acquired by spirit leveling method in Taiwan nowadays. It is worthy, from either academic or practical point of view in engineering survey, to study if GPS-RTK technology can be applied to fast get accurate enough elevation data. The key issue here is geoidal undulation. Most domestic researches of geoid model cover the whole Taiwan area by using gravity surveying data which is relatively difficult to be obtained, time-consuming and much labored while the results are not fit well to some local terrain. In this research all GPS control points with known first-order orthometric heights as well as ellipsoid heights obtained by RTK in Taichung city are used while plane fitting and second-order curve surface fitting methods are also applied to calculate the local geoidal undulation model. In order to get the best solution, issues like data integrality and localization are also considered by changing the quantity and distribution of reference points. At the same time, both randomly selected points and reference points are tested by Support Vector Machine（SVM）method to accurately determine the local geoidal undulation model. The results show that local geoidal undulation models built by each of three methods above meet accuracy requirements of engineering surveying standards ±5 cm. This study provides a fast practical method in getting orthometric heights, meanwhile it can also be used as an academic reference for local geoid model researches.
|
|---|
