Research On The Algorithm For Ground-based Infrared Remote Sensing To Retrieve Land Surface Temperature

Land surface temperature(LST)is an important parameter in the earth science research.And it has an important impact on the growth of vegetation,climate change and human production and life.At the same time,the observation of surface temperature is also one of the important elements of the ground meteorological observation business.At present,there are three main methods of ground surface temperature observation: glass liquid thermometer,platinum resistance ground temperature sensor,and infrared radiation temperature measurement.The meteorological station in China mainly uses platinum resistance to observe the surface temperature now.However,the use of platinum resistance for geothermal measurement is easily affected by multiple environmental factors,and the measurement error is large.Infrared radiation temperature measurement is a new and effective measurement method in the field of surface temperature measurement.It is based on the measurement of surface temperature.And the measurement value of it is more representative.Besides,it is not restricted by the underlying surface materials,so it has a good prospect of application.The main advantage of infrared radiation temperature measurement is high sensitivity and quick responsiveness,and it will not cause damage to the surface heat balance.But the use of infrared measurement of surface temperature needs to consider the influence of emissivity and atmospheric radiation.Andthe error may occur easily in the measurement process.This paper mainly researches on the infrared temperature measurement.The influence of various factors on the accuracy of temperature measurement in the infrared temperature measurement system arecomprehensive analyzed.Based on the true value of the surface temperature measured by the platinum resistance sensor,the inversion calculation method of the ground surface temperature by infrared remote sensing is studied.Then the numerical inversion model for infrared remote sensing of land surface temperature is established,which makes the retrieved infrared surface temperature closer to the temperature measured by platinum resistance and improves the accuracy of infrared measurement.In this paper,a software and hardware platform for data acquisition and analysis of surface temperature observation data of three different underlying surfaces of bare land,grass surface and cement ground is designed and built.The principle and method of measuring surface temperature by infrared radiation and platinum resistance are analyzed.Experimental data of infrared and platinum resistance ground temperatureobservation in June 2017 are selected to compare and analyze.The results show that the difference in temperature measured by different ways in the daytime is larger than that at night for each underlying surface.The reason is that in the daytime,the platinum resistance is half exposed to the surface and is directly radiated by the sun,making the measurement temperature higher.And there is no direct sunlight at night,so the temperature measured by the platinum resistance is close to the surface radiation temperature.Besides,the daily average temperature of the cement surface is the highest,followed by bare land,and the grass surface temperature is the lowest.In order to improve the accuracy of infrared measurement of surface temperature,In this paper,based on Stepan Boltzmann's law,an algorithm for the inversion of surface temperature on the basis of the full radiation temperature measurement is proposed.And in order to further improve the accuracy of inversion,the regression model is used to further modify the existing inversion algorithm.The results show that under the three underlying surfaces,the infrared surface temperature corrected by the regression model is closer to the surface temperature measured by platinum resistance,which achieves the effect of improving the measurement accuracy.Finally,in order to determine the solar radiation is an important influence on the accuracy of LST retrieval,this paper uses the BP neural network inversion algorithm,including two cases of whether or not to engrave the solar radiation at the same time.And finally the influence of the solar radiation on the accuracy of infrared inversion of surface temperature determined.The two methods of inverse calculation of infrared surface temperature proposed in this paper have achieved the intended purpose of the project.It can achieve continuous and reliable observation of land surface temperature on different underlying surfaces.And it provides a basisfor the establishment of infrared remote sensing surface temperature observation method adapted to the needs of meteorological services as well as the business application promotion.