利用静止气象卫星数据提取沙尘信息

根据静止气象卫星通道少且集中在中红外-热红外谱区的特点,将线性光谱混合模型引入沙尘提取研究中.利用相关通道的累积直方图实现端元的自动提取,结合混合像元分解结果与亮温差阈值,建立了针对风云二号C,D星遥感数据的沙尘分布信息提取模型.结果表明:通过原始数据DN值的累积直方图进行端元自动提取,混合像元分解的结果能够实现沙尘信息增强;结合沙尘在11 μm通道与12 μm通道的亮温差为负及在3.7 μm通道与11 μm通道亮温差高于其他目标的特点,对"伪"沙尘像元进行剔除;该方法应用于2005年4月27～28日我国北方地区一次强沙尘暴过程的结果显示,该方法提取沙尘精度较高,能够客观反映沙尘运移过程.

Dust Detection Based on Geostationary Meteorological Satellite Data

Based on the facts that geostationary meteorological satellites have a few channels and most of them are in the middle and thermal infrared region, the linear spectral mixture model (LSMM) was applied to enhance dust information based on automatic selection of end-members according to the accumulation histogram of related channels. Then a hybrid model for deriving dust pixels from FY-2C/2D data was proposed. The results show that the dust infor-mation can be enhanced through LSMM and the endmembers can be derived based on the accu-mulation histogram of the original images automatically. The pseudo dust pixels can be elimi-nated following two criterions. 1) the difference of brightness temperatures in channel 11 μm and channel 12 μm is negative; 2) dust pixels have greater difference between brightness tem-peratures in channel 3.7 μm and channel 11μm than other objects. FY-2C images were utilized to monitor a dust storm process from April 27 to April 28, 2005 in northern China. The appli-cation indicates that the method obtained high accuracy and the transportation process of the dust storm was described quantitatively.