【作者】 费增坪； 王洪庆； 张焱； 郑永光； 刘家峻；

【Author】 Fei Zengping~(1)) Zhang Yan~(2)) Wang Hongqing~(2)) Zheng Yongguang~(3)) Liu Jiajun~(1)) 1)(No 61741 Unit of PLA,Beijing 100081) 2)(Department of Atmospheres Science,School of Physics,Peking University,Beijing 100871) 3)(National Meteorological Center,Beijing 100081)

【机构】 中国人民解放军61741部队； 北京大学物理学院大气科学系； 国家气象中心；

【摘要】 由于缺乏成熟的MCS自动追踪工具,制约了基于静止卫星红外云图的MCS研究的深度和广度。手工目测法识别追踪MCS,工作量大且容易造成误差,无法用于长时间序列卫星云图的MCS普查。本文结合我国中尺度对流天气的特点,在重新定义MCS云团识别标准的基础上,给出了一种基于图像处理和时间序列分析技术的MCS自动识别、存储、追踪和序列特征统计方法。通过2003年淮河大水期间连续时次卫星云图的MCS自动识别、追踪和特征统计,对该方法的应用效果进行了检验。更多还原

【Abstract】 Because have not mature calculate way software for MCS Tracks Automatically the depth and wide are limited of research of MCS based on geostationary satellite infrared imagery,and the result perhaps have more error on objectivity and fixed amount.First,combine the characteristics of the mesoscale convection weather in China and the resolution of time and space of satellite image data,redefine MCS cloud regiment to identify standard.Secondly,According to this standard,design and developped the MCS target to identify and track a method automatically.Finally, though the test of period continue 10 hours of the satellite image during inundation over the Huaihe River region in 2003,the result reveal that the method of MCS identifies and tracks automatically is better.更多还原