抗差趋势面探测海洋磁力测量异常数据

Abnormal Data Detection of Marine Geomagnetic Survey in Robust Trend Surface

数据质量控制是基于多面函数建立海洋局域地磁场模型中的重要环节。针对趋势面检测在异常数据较多时存在一定的局限性,引入了抗差趋势面来检测异常数据。实验显示检测效果优于普通趋势面检测,且在异常数据含量增大时仍有较好的稳定性。进一步比较了基于两种检测方法建立的多面函数模型,结果表明基于抗差趋势面质量控制建立的模型确实具有较好的精度和可靠性。

Controlling the quality of data used for construct regional geomagnetic field model based on multi-surface function is an important section. As for the limitation when excessive outliers existed in the trend surface detection method, the robust trend surface was introduced to detect outIiers. Compared with the trend surface method, the experiment result showed the robust trend surface had a better and steady performance especially with higher outlier ratio. Finally, building the multi-surface function model based on the two methods, the result showed that it was a better method in accuracy and stability of data quality based on robust trend surface.