The Quantitative Relationship Between Land Surface Temperature And Vegetation Indices In Different Regions Based Oil TM Images

With the rapid development of remote sensing technology including thermalinfrared remote sensing technology, which widely used in many fields, our societyincreases the demand of remote sensing applications. Different regions would choosesuitable inversion algorithm to retrieval land surface temperature. Therefore it ismeaningful to discuss the choice of inversion algorithm in different districts. Asimportant parameters about vegetation cover and surface energy exchange, vegetationindex and surface temperature contain a wealth of information on the surface.Research on the relationship between surface temperature and vegetation index hasbeen expanded in multiple regions, so exploring the relationship in different areas canenhance the cognition of changes in surface environmental processes.This paper selects Lansat-5TM images, which cover some areas inShijiazhuang City, Nanchang City and Guangzhou City (some area in Foshan City andDongguan City), and the data sets of ground weather stations. Process and analysedata using geography sofwares, like ENVI and ArcGIS. After the sensitivity analysisfor three retrieval methods of landsurface, compare the results with weather stationdata in that day. Using different methods to sample in three study districts, analyse therelationship between land surface temperature and vegetation indice.Research shows that:(1) In the three study area, considering the distirbution of retireval temperature,the compairson of difference in temperature, comparison with the data of weatherstation and refer to the linear relationship of LST-NDVI, concluded that: theMomo-Window algorithm is suitable for Shijiahzuang study area, the Single-Chanelalgoirth is suitable for Nanchang area, the Momo-Window algorithm is suitable for Shijiahzuang study area. The Momo-Window algorithm has more stable inversionresults in three regions, but with relative higher retrieval temperature in Nanchangarea. The Single-Chanel algorithm gets general high temperature compared with otheralgorithm, and higher temperature than the logical surface temperature with the dataof weather sation. The Image-Based algorithm gets lower temperature among threealgorithm, and the result has small changes.(2)Using the method based on land cover types and based on the interval ofNDVI, and integrated method, concluded that: there are varying degrees of negativecorrelation in three study area, and the oder from significant linear relationship tonon-obvious linear relationship is that: Guangzhou area, Shijiazhuang area, Nanchangarea. The relationship of LST-NDVI is similar with the method based on land covertypes and based on the interval of NDVI. In the integrated method，the correlation inGuangzhou area reduces compared with the former, but still be signiifcant inShijiazhuang area. This is probably caused by the samples (NDVI=0.1) in townscategory and also in the interval0.1-0.2.(3)Single-Chanel algorithm and Mono-Window algorithm have differentsensitivity to the parameters. Momo-Window algorithm has great sensitivity to theatmospheric water vapor, on the comtrary, has smaller sensitivity. Single-Chanelalgoirthm and Mono-Window algorithm have enhanced sensitivity with theatmospheric water vapor increases. Single-Chanel algorithm has higher sensitivitythan Mono-Window algorithm. So pruden use of Single-Chanel algorithm withestimated water vapor.