基于MODIS数据的组分温度反演研究

探讨了一种利用MODIS卫星遥感数据提取组分温度的实用方法．结果表明：对于不同覆盖度的草场，植被的结构较为简单，阴影影响小，在忽略混合像元空间结构特征的前提下，可近似认为混合像元的辐射总能量是各种组分物质辐射能量与其所占面积百分数的加权和，因此，采用约束优化算法——遗传算法进行混合像元分解，用MODIS数据的热红外波段反演基于亚像元尺度的植被／土壤组分温度，不但拓宽了线性光谱混合模式的应用，也可以为地表能量平衡、蒸散研究提供更精确的参数．

Inversion of Component Temperatures Based on MODIS Data

A practical method was developed for component temperatures retrieval with MODIS satellite data. The results show that, to grass lands with different coverage conditions, the radiation is less sensitive to the vegetation shade considering the simple vegetation structure. Under the premise ignoring the space structure of mixed-pixel, the total radiation of mixed-pixel approximately is the weighted sum of each component radiation and its area percentage. Thus, a mixed-pixel can be decomposed with constrained optimization algorithm-genetic algorithm. Vegetation/soil component temperature based on the sub-pixel scale is inversed using thermal infrared bands of MODIS data. It not only broadens the application of linear spectral mixing model, but also provides more precise parameters for estimation of land surface energy balance and evapotranspiration.