Accurate Geometric Correction by using Centroid of Polygon as Control Point

Abstract

Recently, high resolution satellite imageries (about 1 meter spatial resolution) are used in many fields. Generally, they are expected to be overlaid with GIS data and updating existing map. It is important issue for remote sensing to detect land-use/land-cover changes using high resolution satellite imageries. Therefore, the accuracy of geometric correction should be reduced to sub-pixel error. It is necessary to prepare accurate control points to correct geometrically. The control points are Ground Control Point (GCP) collected from the ground surface and corresponded Image Control Point (ICP) from selected imagery. Nowadays, the defined centroids of circles on the ground are used as accurate control points. However, it is difficult to prepare circle type GCPs. The existing types could be evaluated at first part of this study. Continuously, accurate geometric correction could be established using centroids of the polygons as control points. The result showed 0.54m pixel error by applying proposed methodology. It showed that the accuracy of proposed method is approximately same as circle type centroid method. Therefore centroid of rectangular type and polygon type were efficient for the geometric correction of high resolution imageries.