Comparison of gap filling methods in Landsat 7 ETM+ images to estimate crop coefficient

Landsat 7 ETM data is widely used in studies of the spatial distribution Kc and vegetation cover parameters in regional and global scales but SLC failure has greatly reduces its usefulness. Additionally, the failure is permanent and has failed subsequent attempts to recover the SLC, so required and practical way to address this problem is filling the pixels of missed data in the SLC-off images. Although, there are several proposed methods to fill the gap, but still have filled images quality in heterogeneous area is not satisfactory for more applications. This study was conducted to compare the geostatistics and MODIS auxiliary data methods to fill the pixels of missed data in the SLC-off images. The results showed that the IDW method with NRMSE 6.09% was the best method. The fusion with auxiliary images (MODIS) and ordinary Kriging methods resulted in NRMSE 14.75 and 16.9, respectively. The method of fusion with classified auxiliary images (MODIS) presented the lowest accuracy in estimating missed data.