Accuracy analysis of polynomial rectification for Quickbird imagery

Abstract

Geometric rectification is a process to relate space coordinate systems of satellite imagery with the ground coordinate systems. Rectification involves georeferencing in which ground coordinates are assign to image data. Photogrammetric data extraction from satellite imagery can be done after the image being geometrically corrected. There are several rectification methods proposed by photogrammetric remote sensing experts. Each method is having its own characteristics in term of mathematical model used, geometrical accuracy and computation constrains. Conventional polynomial rectification was reported as not suitable for geometric rectification of high resolution satellite imagery. The study is conducted to access and evaluate the accuracy of polynomial rectification for QuickBird imagery. Analysis will be focused to the effect of 1st, 2nd and 3rd polynomial order on rectified imagery. This paper briefly reviews polynomial rectification model and presented the result of the study. Sub-pixel accuracy can be achieved for fairly flat area.