本文研究的是多颗气象卫星的多信息源的数据同化问题，主要应用于城市温度分布状况的监测系统，在使用多源数据前，首先需要对数据进行同化处理。同化的方法是对不同卫星探测辐射值进行直接对比，然后进行投影变换几何粗校正和图像平移精校正，使多源数据在几何位置上能得以高精度的配准。设计了软件流程，编写了ActiveX 组件，同化工作在服务器端完成，使用ODBC 对数据库存取，最后利用实际遥感数据对系统进行了检验，获得的结果符合反演使用的要求。关键词: 遥感信息处理；数据同化；极轨气象卫星；多源数据Designing for Thermal Infrared Data Assimilation of Meteorological Satellite CAO Jian-qing1,Hu Yong2,ZHU Shan-you2,YIN Qiu2, GONG Cai-lan2 (1.Shanghai Second Polytechnic University, Shanghai 201209,P.R.China 2. Shanghai Institute of Technical Physics, Shanghai 200083, P.R.China) Abstract: Assimilation of different satellite data is researched , it applies to the monitoring system of the urban temperature. The thermal infrared data from polar orbit meteorological satellites must be assimilated before they are used. The method of assimilation is direct comparison of radiance detected by different satellites, geometric rough correction by projection transform and precious correction by image transfer, it can register multi-sources data in geometric positions with higher precision. The procedure of the software is designed ,ActiveX component is programmed. Assimilation code is executed in the server , accessing the database by ODBC. The system is tested by inputting actual remote sensing data finally, the result obtained accords with the demand.Key words: Remote sensing information process; Data assimilation; Polar orbit meteorological satellite;Multi-sources data