Accuracy evaluation orthorectification of SPOT image

Abstract

Orthorectification in the study conducted on 14 scene multispectral imagery SPOT4 SPOT2 and West Java. The composition of the scene to form a wide block with no gaps, making it possible to do block triangulation. This triangulation is used to determine the relationship between the satellite image and the earth. The information generated from the triangulation will be used as input during orthorectification. Ground control points (GCP) were taken from the data vector Indonesian Topographic Map 1:25,000 scale and elevation data extracted from the DEM derived from contour maps RBI data is the same. Triangulation blocks performed with various configurations GCP and tie points (TP), the minimal configuration, just along perimeter blocks, regularly spaced regularly spaced half scene and a third scene. Each configuration is lowered into several configurations based on the use of TP. Triangulation accuracy result was tested with 244 check points (CP) that spread in the block configuration can produce a minimum GCP accuracy RMSE (Root Mean Square Error) 1.25 pixels if equipped with a TP meeting. GCP configuration on the perimeter of the block is not recommended for use if it is not equipped with TP. Highest accuracy resulting from configuration GCP/TP per a third scene with RMSE 0.98 pixels. Accuracy will be determined by the distribution and the number of GCP and the use of TP. Best accuracy of this research is suitable for medium-scale map updates 1:65.000.