中文摘要：

本文从分析航空重力向下延拓过程中偶然误差和系统误差的变化特性入手,进而提出处理办法。首先,利用试验说明移去恢复法局限性,同时表明需处理系统误差和偶然误差的必要性。然后,采用理论推演和数值模拟计算分别估计了系统误差和偶然误差影响,试验结果发现：系统误差影响和偶然误差影响均与数据格网间隔、向下延拓高度呈线性关系,当格网化间隔较小和延拓高度较高时系统误差影响和偶然误差影响较大。最后,提出使用半参数模型和正则化算法的两步法估计系统误差和减弱偶然误差影响,试验结果说明两步法处理向下延拓各类误差影响优于仅用半参数模型或正则化算法的结果,在试验数据的偶然误差标准差为2×10~（-5） m/s~2、恒值系统误差3×10~（-5） m/s~2和变值系统误差标准差约1.3×10~（-5） m/s~2时,以及向下延拓高度6.3 km和格网间隔6′的条件下,两步法向下延拓结果的精度可达2.3×10~（-5） m/s~2。

英文摘要：

The variation characteristics of the random errors and systematic errors of airborne gravity （AG） are analyzed in the downward continuation （DWC） process, and then specific processing methods are presented to deal with the influence of systematic errors and random errors.Firstly, the limitation of remove-compute-recover （RCR） was generated from the RCR experiment, and it is necessary to deal with random errors and systematic errors.Systematic error effects and random error effects were calculated based on theoretical deductions and numerical simulations.The results showed the linear relationships between systematic error effects or random errors effects and grid spaces or DWC heights.It was concluded that the smaller grid spaces and higher DWC heights would increase the systematic error effects and random error effects.Ultimately, the two-step method of semi-parametric model and regularization method were proposed to estimate systematic error and to weaken random error effects.And the experimental results showed the two-step method was more effective dealing with DWC error effects than semi-parametric model and regularization method separately.Specifically, conditions of experiment were that standard deviation of random error was 2×10-5m/s2, bias 3×10-5 m/s2 and the standard deviation of variable systematic error was about 1.3×10-5m/s2, the DWC height was 6.3km and the resolution was 6′.And two-step method experimental accuracy could reach about 2.3×10-5m/s2.