Retrieving Land surface Component Temperature UsingMulti-Angle Thermal Infrared Data

摘要

As non-isothermal mixed pixel is widely existed,the pixel-mean temperature cannot adequately represent theactual thermal state of land surface. The row crop was chosenas target to discuss the problem of component temperatureretrieval. At first, the matrix model was found to express thethermal radiant directionality of the target. Then correlation ofmulti-angle infrared radiance was analyzed. In order toincrease the retrieving accuracy, we chose the retrievableparameters and established the iterative methodcombining with inverse matrix to retrieve componenttemperature. It was proved by field experiment that themethod could improve the retrieving accuracy andstability remarkably.