摘要:本文将计算机图形学骨架概念应用到气象学领域，发展了回波图像预处理、骨架修剪处理以及长宽比量化处理技术，该方法能自动识别出雷达回波拼图中符合气象学标准的线状对流系统（quasi-linear convective systems，QLCSs）。首先结合2016年黄淮地区一次双QLCSs过程给出了基于骨架的QLCSs客观量化算法的具体技术流程，然后利用该方法对2016年6月安徽地区的QLCSs进行客观筛选，并进一步量化识别QLCSs的移动特征，结合灾害天气实况与主观识别进行对比评估，结果表明：结合气象学标准改造的骨架图像识别算法，较好保留了气象回波形状信息，在准确量化对流系统长短轴的基础上，实现QLCSs的有效识别。而获得的量化移动矢量等特征，一方面可应用于致灾QLCSs的分类研究，为开展长序列统计及致灾机理分析提供个例识别方法和量化特征，另一方面也为QLCSs的短临监测预警业务提供新的思路。
关键词:线状对流系统/
图像识别/
骨架/
雷达回波/
量化特征
Abstract:In this paper, we apply the skeleton concept from the field of computer graphics to meteorology and the development of echo image preprocessing, skeleton pruning, and length-width ratio quantization techniques. The quasi-linear convection systems (QLCSs) in radar echo mosaics that conform to meteorological standards can automatically be identified using this method. First, we introduce in detail the identification algorithm based on a double QLCSs process in the Huang–Huai area in 2016. Then, we use this method to objectively identify the QLCSs in the province of Anhui in June 2016 and quantify the moving characteristics of the QLCSs. We then compare weather disaster observations with their subjective identification. The results show that information about the shape of meteorological echoes is well preserved using the skeleton image identification algorithm and effective identification of QLCSs is realized base on the accurate quantification of the long and short axes of the convection system. Quantitative characteristics such as moving vectors can be applied to the classification of QLCS disasters to provide an identification algorithm and quantitative features for long–series statistics as well as an analysis mechanism for weather disasters. In addition, this concept provides a new technique for monitoring and warning by QLCSs.
Key words:Quasi-linear convective systems/
Image identification/
Skeleton/
Radar echo/
Quantitative features
PDF全文下载地址:
http://www.iapjournals.ac.cn/dqkx/article/exportPdf?id=5d0b86a6-db0d-48dc-a35b-814334da5845
