高精度重复相对重力观测量提取

采用RANSAC算法剔除观测数据中的离群值，再使用线性内插法进行补全，利用整体投影计算的思想提取两点间的相对重力值，并对其精度和标准差进行检验。结果表明，动态重力观测的残差最大值为4.641 μGal，重复性标准差最大值为4.384 μGal，均优于5 μGal。该方法可获得较高精度的重力观测数据，为在复杂环境下获取相对重力值提供一种新方法。

Extraction of High-Precision Repeated Relative Gravity Observation Value

In this paper, we eliminate outliers in observation data by RANSAC algorithm, and the linear interpolation method is used to complete the data set. The relative gravity value between two points is extracted by overall projection calculation, and its accuracy and standard deviation are tested. The results show that the maximum residual value of dynamic gravity observation is 4.641 μGal, and the maximum repeatable standard deviation is 4.384 μGal, both less than 5 μGal. This method can obtain high-precision gravity observation data, and provides a new method for obtaining relative gravity values in complex environments.