Improving geometric construction of high resolution SAR images using Kriging-based surrogate modelling in mountainous terrain of Malaysia

The spatial resolution of SAR images has been significantly improved in recent years. Conventional geometric correction models are computationally expensive, and require a significant amount of time to process high-resolution SAR data. This study concentrates on the Kriging-based surrogate algorithm to improve a rigorous geometric correction model for high-resolution SAR images. This study exploits a global digital elevation model (DEM) to build a 10-m grid of elevation points. The Kriging-based geometric correction model estimates the elevation at unmeasured locations between the grid points to reconstruct the geometry of high-resolution SAR images. The implemented method is a robust approach to improving the geometric construction of SAR images in mountainous terrains. The obtained results show that the Kriging-based geometric correction model enhances the computation efficiency of rigorous physical geometric models and improves the geometric construction of SAR images. The geometrical error of SAR pixels is found to be between 18 and 169 pixels through 92 selected ground control points. In contrast, the Kriging-based geometric correction model has reduced the geometrical error of SAR pixels by up to 1 pixel (i.e. 6 m) with significantly less computational redundancy.