许健民,郭强,陆其峰,陆风,张晓虎. 2014. 风云气象卫星数据处理算法的若干创新[J]. 气象学报, 72(5):1023-1038, doi:10.11676/qxxb2014.086
风云气象卫星数据处理算法的若干创新
Innovations in the data processing algorithm for Chinese FY meteorological satellites
投稿时间:2014-01-29  修订日期:2014-07-30
DOI:10.11676/qxxb2014.086
中文关键词:  气象卫星  数据处理算法  定位  定标  数据同化
英文关键词:Meteorological satellite  Data processing algorithm  Image navigation  Radiation calibration  Data assimilation
基金项目:国家自然科学基金项目(40275007、41275036、40971200、41075019、41275034、91338203、40705037)、财政部气象公益专项(GYHY201206002、201306001)、科技部(863-2003AA133050、2012AA120903)。
作者单位E-mail
许健民 国家卫星气象中心, 北京, 100081  
郭强 国家卫星气象中心, 北京, 100081 guoqiang@cma.gov.cn 
陆其峰 国家卫星气象中心, 北京, 100081  
陆风 国家卫星气象中心, 北京, 100081  
张晓虎 国家卫星气象中心, 北京, 100081  
摘要点击次数: 3671
全文下载次数: 2681
中文摘要:
      在图像定位、辐射定标、数据同化三个方面各选择一项工作,简述风云气象卫星(FY)数据处理算法的创新。在自旋稳定静止气象卫星观测图像上,地球圆盘中心行序号的时间序列里,存在关于卫星自旋轴指向的信息,用其可以求解出卫星的姿态参数,并确定地球圆盘中心的南北位置。卫星每自旋一周,可以看到太阳和地球各一次。只要卫星的位置和姿态已知,在卫星上看太阳与地球的夹角即可精确求出并预报,可确定地球圆盘中心的东西位置。在此基础上推出风云二号(FY-2)图像定位方程和算法。用这种算法,风云二号气象卫星图像定位的精度达到了像元级。工作在地球静止轨道上的风云二号气象卫星,接收到的太阳辐射存在显著的日变化和年变化,这就决定了其3.5 μm以上热发射波段的辐射响应,必然存在从小时到年际等不同时间尺度的周期性变化,需要精确标定。一方面,利用星载内黑体观测结果,在精确修正前光路辐射贡献的基础上,可以得到小时级辐射定标参数的变化特征;另一方面,利用在轨观测到的月表辐射为基准,可以修正内黑体定标结果在数日至年际尺度上的系统性偏差。采用这些定标算法后,风云二号气象卫星红外辐射定标精度优于1 K。卫星遥感仪器在轨性能参数对数据质量起着重要作用,这些参数可能受实验室条件所限测不准,也有可能上星后受空间环境影响发生了改变。建立了一个极轨气象卫星遥感仪器在轨性能参数变分优化模型(SIPOn-Opt模型),通过观测约束优化获得更真实的仪器特征参数。将该模型应用于风云三号气象卫星(FY-3)大气探测仪器后,使其卫星数据与欧美同类先进遥感仪器数据质量相当,并对数值天气预报准确度有改进。
英文摘要:
      This study introduces some innovations in the data processing algorithm for Chinese FY meteorological satellites. Issues about satellite image navigation, radiation calibration, and data assimilation are discussed. A time series of the earth’s disk center-line count provides information on the orientation of the satellite spin axis. With this information, the altitude parameters of the satellite and then the earth disk location in the south-north direction may be solved. In each spin cycle, the satellite views the sun and the earth. Given the satellite position and altitude, the angle (β) subtended at the satellite by the sun and the earth can be calculated and predicted. Thus, the earth’s disk location in the east-west direction is fixed. Based on this principle, we derived an automatic image navigation algorithm for FY-2 geosynchronous meteorological satellites with an accuracy approaching pixel level. The FY-2 meteorological satellite traveling in a geostationary orbit suffers a large amount of radiation from the sun. The radiation varies on both diurnal and annual scales, which causes radiation responses in the thermal infrared (IR) bands wherein the wavelengths greater than 3.5 μm vibrate periodically on scales of hours to years. These vibrations must be precisely calibrated. First, based on the accurate estimation of the radiant contribution from the front-optics, the variation characteristics of the calibration parameters are obtained on a temporal scale of hours from the space-borne inner-blackbody (IBB) measurement results. Second, the in-orbit measured radiation of the lunar surface is referenced and utilized to correct the systematic bias of the IBB calibration from daily to annual scales. By using such algorithms, we achieved a calibration accuracy of the FY-2 satellite’s IR imagery of less than 1 K. The on-orbit satellite instrument parameters play an important role in data quality; however, they may be mis-measured due to limitations in the measurement conditions or may be changed due to the space environment after launch. A satellite instrument parameters on-orbit optimizer (SIPOn-Opt) for a polar orbit meteorological satellite was developed to optimize the true state of the instrument parameters on-orbit with regard to the observation constraints. When applying the SIPOn-Opt to FY-3 sounding instruments, the FY-3 data quality was much improved, compared to its European and the U.S. polar orbit meteorological satellite counterparts, leading to improved forecast skill of numerical weather prediction.
HTML   查看全文   查看/发表评论  下载PDF阅读器
分享按钮