Research on fast geometric precision correction of HJ-1 remote sensing image

As the basis application of remote sensing image ,the satellite remote sensing image geometric precision correction is an important means that to weaken the image itself with the real terrain morphological differences. In this paper, we used HJ-1 satellite CCD data as the research object, using the SIFT algorithm in combination with GPU parallel computing fast extracting and matching feature points, based on polynomial correction model to build calibration system, independent of the HJ-1 satellite imagery geometric precision correction, and achieved a higher accuracy.