论文部分内容阅读
考虑到高光谱大气红外探测器通道之间的相关性、变分同化的时效性等,需要进行通道选择。利用主成分—逐步回归法进行AIRS通道选择研究。由于短波CO2通道易受太阳光影响,分白天和夜晚进行。具体执行过程中,首先进行通道预处理,然后分别对温度和湿度雅可比矩阵进行主成分分析,并采用逐步回归法找出对前几个主成分影响较大的通道得到入选的通道子集。进而根据经验和实际观测资料,为了达到全局最优并兼顾局部,基于分区的思想,采用主成分—双区逐步回归法进行通道选择。结果表明:1利用AIRS进行反演时,合理选择通道是非常必要的;2主成分—双区逐步回归法得到的通道组合进行温度、湿度反演的误差整体比基于信息熵分步迭代小。
Taking into account the hyperspectral atmospheric infrared detector channel correlation between the variability of the timeliness of assimilation, the need for channel selection. AIRS CHANNEL SELECTION RESEARCH BASED ON PRINCIPAL COMPONENT - STEP - REGRESSION METHOD. As the short-wave CO2 channel susceptible to sunlight, sub-day and night. In the concrete implementation process, the channel preprocessing is performed first, and then the principal components analysis is performed on the Jacobian matrix of temperature and humidity respectively, and the stepwise regression method is used to find out the channel subsets that have a great influence on the first few principal components. Furthermore, based on the empirical and practical observation data, in order to achieve the global optimum and take into account the local and district-based thinking, the principal component-two-step stepwise regression method is used to select the channel. The results show that: 1 When using AIRS for inversion, it is very necessary to select the channel reasonably. The error of temperature and humidity inversion for channel combination obtained by 2-component stepwise regression method is smaller than that based on information entropy step by step.