针对靶场复杂的试验环境及测量数据的特点，深入分析了异常数据产生的原因，提出了一种利用多项式拟合对测量结果数据进行双向检验，剔除异常数据的方法，并在检验结果满足修复条件时，对剔除数据进行加权修复。实际数据测试表明:本方法能够有效克服数据突变和段落性阶跃等干扰因素的影响，具有较好的模型稳定性和较高的异常数据剔除率，满足试验任务实际需求，具有较高的工程应用价值。

In view of the complex test environment and characteristics of measurement datathrough a deep analysis of the causes of abnormal dataa method for eliminating abnormal data based on two-sided test of measurement result data by use of polynomial fitting was presented.When the test result met the repairing conditionweighted repairing was conducted for the eliminated data.Real data test showed that this method can effectively overcome the influence of data mutation and paragraph stephas better model stability and higher abnormal data eliminating ratewhich can meet the actual requirement of experimental task and is of high practical value in engineering.