由J.Vondrak在1967年提出的Vondrak滤波算法，发展已逐渐成熟，其可以在观测数据的变化趋势、拟合函数均未知的前提下，对数据资料进行滤波平滑.基于Vondrak滤波，又发展出了Vondrak-Cepek滤波，其可以在保证观测数据序列长期特性的前提下结合其一阶导数的短期性能进行滤波平滑.Vondrak滤波与V-C滤波基于其特性，在各类研究中都得到广泛应用，其中较为常用的是时间尺度领域.

但由于环境、设备、人为等各方面影响，观测数据中可能会存在粗差，而粗差会导致滤波值发生偏移，从而大大降低滤波精度. 然而随机抽样一致性算法可以从观测数据中分离出粗差点，并进行剔除. 因此本文在原始Vondrak滤波的基础上，结合了随机抽样一致性算法，将观测数据中的粗差进行识别与剔除，提高滤波精度，同时对比V-C滤波与随机抽样一致性算法的结合，比较二者精度. 文章主要内容如下：

通过研究学习Vondrak滤波算法、V-C滤波算法的基本原理与推导过程，掌握其滤波方法，再将随机抽样一致性算法分别与上述两者进行结合改进，将改进后的滤波算法用于模拟算例来进行比较，并调整算法参数，最后分别将改进前的算法与改进后的算法应用于实际时间、频率数据，对比其滤波结果，证明随机抽样一致性可以提高滤波精度.

The Vondrak filtering algorithm proposed by J. Vondrak in 1967 has gradually matured in its development. It can filter and smooth data without knowing the changing trend and fitting function of the observed data. Based on Vondrak filtering, Vondrak Cepek filtering has also been developed, It can be combined with the short-term performance of its first derivative for filtering and smoothing while ensuring the long-term characteristics of the observed data series. Vondrak filtering and V-C filtering are widely used in various studies based on their characteristics, with the most commonly used being in the field of time scales.

However, due to various factors such as environment, equipment, and human factors, there may be gross errors in observation data, However, gross errors can cause the filtering values to shift, greatly reducing the filtering accuracy. However, the random sampling consistency algorithm can separate gross errors from the observation data and eliminate them. Therefore, this article combines the original V-C filtering with the Randon sample consensus to identify and eliminate gross errors in the observation data, improving the filtering accuracy. At the same time, it compares the Vondrak filtering with the Randon sample consensus, Compare the accuracy of the two. The main content of the article is as follows:            

By studying and learning the basic principles and derivation process of Vondrak filtering algorithm and V-C filtering algorithm, one can master their filtering methods. Then, the Randon sample consensus is combined and improved with the above two algorithms, and the improved filtering algorithm is used for simulation examples to compare and adjust algorithm parameters.  Finally, the improved algorithm and the improved algorithm are applied to actual data to compare their filtering results, Proving the Randon sample consensus can improve filtering accuracy.