A Point-Scatterer Approach to Terrain Reconstruction for Imaging Radar Simulation

Abstract

We present a point-scatterer approach for simulating imaging radar systems over rugged, heterogeneous terrain. As airborne synthetic aperture radar (SAR), MIMO systems, and UAV navigation algorithms become more common, the need for high-fidelity simulation environments increases. Analytically calculating electromagnetic reflections from complex landscapes with diverse materials is computationally heavy. To solve this, we propose an algorithm that converts a geometric terrain model into an electromagnetic one. The process takes a Digital Elevation Model (DEM) and a surface classification matrix to generate a triangular mesh scaled to the radar's spatial resolution. Lineof-Sight (LoS) calculations are used to account for terrain selfshadowing. We treat each visible triangle centroid as a point scatterer, assigning a complex reflection amplitude based on material-specific Radar Cross Section (RCS) models and dynamic incidence angles. The simulation features a complete radar signal chain, covering transmitter/receiver distortions, channel effects, and antenna array topologies. Our results show the successful synthesis of realistic radar images that reproduce geometric artefacts like layover, foreshortening, radar shadows, and speckle noise. Evaluating the structural fidelity of the simulated point clouds using the Iterative Closest Point (ICP) algorithm with Mahalanobis distance confirms their accuracy against heterogeneous errors. This approach provides a computationally viable framework for radar system demonstration.