Abstract:

ETM+ images from Landsat7 exists defects since May 31, 2003.This is caused by the failure of the Scan Lines Corrector on the Enhanced Thematic Mapper.These images called SLCOFF data have some black scan lines which have no data.The lost data is about 25% of whole scene,and this makes them difficult for normal use.However,the data itself still retains some good radiometric and geometric properties;it can be used in some special purposes if they are properly repaired.This paper described how to use the Adaptive Local Regression Algorithm(ALR)to restore these images,and then used the restored image to inverse the parameters of water quality of East Lake in Wuhan,China.The results show that,SLCOF image could be repaired quite well by ALR algorithm,and the repaired image could be used to build quite good empirical retrieval models for water quality parameters such as Chla,SD,TP and TN in East Lake with in situ water quality monitoring data by multiple regression analysis.The square correlation coefficients〖WTBX〗(R〖WTBZ〗2)of the models were 086,075,073 and 071,respectively.The distributions of the retrieval water quality parameters agreed with the actual situation of the lake quite well.The author suggests that ETM+ SLCOFF data shouldn’t be neglected.They have some advantages,such as high spatial resolution,very rich in archived data and can be downloaded for free from NASA's servers.After repaired properly as the paper shows,they can be used as supplement or substitution for other remote sensing data while it is insufficient or unavailable.〖