基于DBSCAN聚类方法的强回波定位追踪  

Strong echo location and tracking based on DBSCAN clustering method

作　　者：徐碧裕 徐加民[2] 于东海[2] 胡丽华[1] 叶朗明[1] Xu Biyu;Xu Jiamin;Yu Donghai;Hu Lihua;Ye Langming(Jiangmen Meteorological Bureau,Jiangmen Guangdong 529000;Meteorological Bureau of Xinhui District,Jiangmen Guangdong 529000)

机构地区：[1]广东省江门市气象局,广东江门529000 [2]江门市新会区气象局,广东新会529000

出　　处：《气象研究与应用》2020年第3期61-64,共4页Journal of Meteorological Research and Application

基　　金：广东省气象局科学技术研究项目(GRMC2019Z04)、江门市气象局科技项目(201901)。

摘　　要：运用PHP网络编程语言和DBSCAN聚类分析方法,提出了一种新的解析应用该资料的强回波定位追踪方法,针对江门蓬江X波段相控阵雷达组合反射率图片资料,统计2020年6月持续104次强回波定位追踪情况以验证此方法的准确率。结果显示,在DBSCAN聚类参数Eps=10,minPts=2设置下,强回波聚类追踪效果较优,各项结果准确率均值达到85.35%,能满足实际业务所需。Based on the high spatial-temporal resolution image data of Pengjiang phased array radar,using PHP network programming language and DBSCAN clustering analysis method,a new strong echo location and tracking method was proposed for the combination reflectivity image data of X-band phased array radar in Pengjiang,Jiangmen.In order to verify the accuracy of this method,104 times of strong echo tracking in June 2020 were counted.The results show that when the DBSCAN clustering parameters EPS=10,minpts=2,the strong echo clustering tracking effect is better,and the average accuracy of strong echo tracking results is 85.35%,which can meet the needs of actual business.

关 键 词：DBSCAN聚类 强天气 雷达监控 自动报警 

分 类 号：P429[天文地球—大气科学及气象学]