顾及剩余地形模型的区域似大地水准面精化OA

To construct a regional high-precision quasi-geoid,this study addresses the in-sufficient spatial resolution of existing gravity field models by compensating for truncation errors using the Residual Terrain Model(RTM).Based on GNSS/leveling data from two distinct test areas,the applicability of high-order gravity field models such as EGM2008 was evaluated,and residual height anomaly fitting models were developed using polynomial surface fitting,multi-surface functions,and BP neural networks.The results indicate that the EGM2008 model performs better in plain areas,whereas the SGG-UGM-2 model yields superior accuracy in mountainous regions.After RTM correction,the average fitting accuracy of height anomalies improved by 11.85%and 8.12%in the two regions,respec-tively.Among the tested methods,the multi-surface function approach achieved the high-est accuracy,with computed precisions of 1.67 and 3.67 cm for the respective regions.These findings demonstrate that incorporating RTM correction effectively enhances the re-finement of regional quasi-geoid models.