Particle Swarm Optimization of RFM for Georeferencing of Satellite Images

Rational function models (RFMs) provide one of the best methods of extracting spatial information from high-resolution satellite images, particularly when sensor parameters cannot be accessed. RFM terms have no physical meaning, and hence, all of these terms are typically used in conventional solutions. As a result, more ground control points (GCPs) are required, and the model is prone to overparameterization. In this letter, a modified particle swarm optimization is applied to identify the optimal terms for RFMs. In comparison to conventional models, experimental results demonstrate how well the proposed algorithm can determine an RFM, which is optimal in both the total number of terms and the positional accuracy. The proposed algorithm is determined to be efficient when subpixel accuracy can be obtained with four GCPs in IKONOS-Geo image.