Landsat 8劈窗算法精度评价及适用性分析

Accuracy Evaluation and Applicability Analysis for Landsat 8 Split Window Algorithms

针对Landsat 8数据的两种地表温度反演算法:Jiménez-Mu?oz (JM)和Offer Rozenstein(OR)劈窗算法,利用辐射传输模型模拟分析算法的绝对精度以及平均绝对精度随参数的变化关系。基于误差传递理论,分析对比两种算法的总精度在此基础上评价两种算法的适用条件。结果表明:(1) JM算法高估地表温度水汽含量较高时算法总体精度较高;(2) OR算法低估地表温度,水汽含量较低时,算法总体精度高于JM算法。因此建议水汽含量较低时使用OR算法水汽含量较高时使用JM算法。

In this paper,for two land surface temperature retrieval algorithms based on Landsat 8 data:Jiménez-Mu oz(JM)and Offer Rozenstein(OR)split-window algorithm,the radiation transmission model MODTRAN was applied to simulate and analyze the absolute accuracy of the algorithm and the relationship between the average absolute accuracy and the parameters.Based on the error transfer theory,the total accuracy of the two algorithms was analyzed and compared.The applicable conditions of the two algorithms were evaluatedbasedon this basis.The results showed that:(1)JM algorithm overestimatedthe surface temperature,the higher water vapor content was,the higher overall accuracy of the algorithm was;(2)OR algorithm underestimated the surface temperature,and when the water vapor content waslow,the overall accuracy of the algorithm washigher than that of the JM algorithm.Therefore,this article suggestedthat OR algorithm should be used when the water vapor content waslow,and JM algorithm should be used when the water vapor content washigh.