基于数据特征的雷达径向干扰回波识别方法

RESEARCH ON AN IDENTIFICATION METHOD FOR RADAR RADIAL JAMMING ECHOES BASED ON DATA CHARACTERISTICS

关键词:  径向干扰  数据特征  新一代天气雷达  径向数据判定法

Key words:  radial interference  data characteristics  next-generation weather radar  radial data determination

摘要:

通过对2020年1月广西新一代天气雷达基本反射率因子拼图产品中存在径向干扰回波的166个样本进行分析，提出一种基于径向数据特征分析的径向干扰回波识别方法：径向数据判定法（Radial Data Determine，RDD）。径向数据判定法方法通过建立4个特征参数来描述反射率因子纵向和横向的数据特征，基于决策树方法识别出径向干扰回波。运用径向数据判定法对该166个样本进行回算，发现径向数据判定法不仅能识别出小幅度的径向干扰回波，而且能有效识别出大幅度高强度的径向干扰回波，识别和滤除效果良好。经过对2019—2020年不同天气过程下的约3?784个样本（包括有或无干扰回波）进行了应用检验，同样发现径向数据判定法对径向干扰回波的正确识别率高，误识别损失小，具有良好的泛化能力和业务应用前景。最后，给出了径向数据判定法应用中存在的问题和未来工作展望。

Abstract:

By analyzing 166 jamming echo samples of Doppler weather radar reflectivity mosaic products for Guangxi Region in January 2020, this work put forward a method, known as Radial Data Determination (RDD), for identifying radial jamming echoes based on radial data characteristics. The RDD method describes the longitudinal and transverse characteristics of the radar reflectivity by establishing four characteristic parameters and uses the decision tree to identify the radial jamming echoes. For clear air echoes of training samples, the RDD method identified not only radial interference echoes at small scales but also those at large scales and high intensity without any preset azimuth threshold. A training test showed that the RDD method had a good effect on identifying and filtering the radial interference echoes. As indicated by the results of an application test for more than 3800 samples (including echoes with clear air or precipitation) with different weather processes from 2019 to 2020, the RDD has a high rate of accurate identification for radial interference echoes and a small loss resulting from erroneous identification, suggesting that it has promising capacity for generalization and good prospect for operational application. In the last part of the work, the issues and outlooks of the RDD are presented.