Potentials and limitations of SAR image simulators – A comparative study of three simulation approaches

Various applications, like for instance algorithm design, mission planning, geo-referencing, change detection , Automatic Target Recognition (ATR) or SAR data analysis, rely on simulated synthetic aperture radar (SAR) images. However, there are different SAR simulation techniques with different advantages and disadvantages. Depending on the needs of a certain application, a suitable SAR simulation technique has to be used. This paper compares three SAR image simulation approaches, RaySAR, CohRaS Ò , and SARViz, showing their similarities and differences. RaySAR and CohRaS Ò are two ray tracing based SAR image sim-ulators. RaySAR is based on the open-source software POV-Ray, while CohRaS Ò is developed as SAR sim-ulator from scratch. The third simulator, SARViz, is a real-time SAR simulator based on the rasterization approach. The geometrical features of the three simulators are compared and the differences and different applications are analyzed. The simulation of synthetic aperture radar (SAR) data has many applications in radar remote sensing, for example in sensor and algorithm design, mission planning, geo-referencing, SAR data analysis and interpretation, etc. This wide variety of applications leads to a variety of SAR simulation approaches, each fitting best for specific objectives and preconditions. However, no simulator is suitable for all scenarios or applications. According to Franceschetti and Migliaccio (1995) two main types of SAR simulation systems can be distinguished: SAR image simulators, which directly provide focused images, and SAR raw signal simulators, which deliver SAR raw data for processing. In this context, different signal reflection models may be utilized. Scattering can for example be simulated using the Kirchhoff physical optics (PO) (see Ulaby et al., 1986), the geometrical optics (GO) approximations (Franceschetti et al., 2003), the integral equation method (IEM) (Xu and Jin, 2006), a more simplified approach related to the Phong shading model (Brunner et al., 2011), or combinations of these. In our work, we focus on SAR image simulators for applications in urban areas. Therefore, we will use the term SAR simulators as short form for SAR image simulators in the remaining text. SAR simulators assist radar remote sensing in urban areas in different ways. Mission planning is, for example, one important application (Gelautz et al., 1998). Another important field is the assistance of SAR image interpretation in complex urban areas, where SAR sim-ulators help to identify the origin of backscattering contributions (Brunner et al., 2011) (Auer, 2011). Furthermore, SAR simulators can be used to create synthetic input data for the …