2D Coordinate Transformation Between Aerial and Satellite Images Using Helmert and Cubic Splines Methods
Mehmet Arkali, Saziye Ozge Atik, Cengizhan Ipbuker

Istanbul Technical University

Abstract

The integration of spatial datasets from various sources necessitates precise coordinate transformations to ensure geometric consistency and interoperability. This study presents a coordinate transformation conducted between a high-resolution Pleiades satellite image and an orthophoto of the Ayaza&#287-a Campus at Istanbul Technical University. A total of 43 conjugate control points were identified within both datasets to effectively model the spatial relationship between the two coordinate systems.Two distinct transformation methodologies were applied: the 2D Helmert similarity transformation and the cubic spline transformation. The Helmert model is a rigid-body transformation, characterized by four parameters: translations along the X and Y axes, a uniform scale factor, and a single rotation angle. Conversely, the cubic spline method utilizes higher-order polynomial functions to address local distortions in the coordinate relationships.The transformation parameters were estimated using least squares adjustment. The results from the Helmert transformation indicated root mean square error (RMSE) values of 1.506 meters in the X direction and 1.353 meters in the Y direction. In contrast, the cubic spline method demonstrated enhanced accuracy, yielding RMSE values of 1.458 meters in the X direction and 1.248 meters in the Y direction. These findings suggest that while the Helmert transformation is effective for global alignment, spline-based approaches are superior for capturing local spatial variations in coordinate systems.

Keywords:

Topic: Topic D: Geospatial Data Integration