Aerial images are important in different scenarios including surface cartography,surveillance, disaster control, height map generation, etc. Synthetic Aperture Radar(SAR) is one way to generate these images even through clouds and in the absence of daylight. For a wide and easy usage of this technology, SAR systems should be small, mounted to Unmanned Aerial Vehicles (UAVs) and process images in real-time. Since UAVs are small and lightweight, more robust (but also more complex) time-domain algorithms are required for good image quality in case of heavy turbulence. Typically the SAR data set size does not allow for ground transmission and processing, while the UAV size does not allow for huge systems and high power consumption to process the data. A small and energy-efficient signal processing system is therefore required. To fill the gap between existing systems that are capable of either high-speed processing or low power consumption, the focus of this thesis is the analysis, design,and implementation of such a system. A survey shows that most architectures either have to high power budgets or too few processing capabilities to match real-time requirements for time-domain-based processing. Therefore, a Field Programmable Gate Array (FPGA) based system is designed, as it allows for high performance and low-power consumption. The Global Backprojection (GBP) is implemented, as it is the standard time-domain-based algorithm which allows for highest image quality at arbitrary trajectories at the complexity of O(N3). To satisfy real-time requirements under all circumstances, the accelerated Fast Factorized Backprojection (FFBP) algorithm with a complexity of O(N2logN) is implemented as well, to allow for a trade-off between image quality and processing time. Additionally, algorithm and design are enhanced to correct the failing assumptions for Frequency Modulated Continuous Wave (FMCW) Radio Detection And Ranging (Radar) data at high velocities. Such sensors offer high-resolution data at considerably low transmit power which is especially interesting for UAVs. A full analysis of all algorithms is carried out, to design a highly utilized architecture for maximum throughput. The process covers the analysis of mathematical steps and approximations for hardware speedup, the analysis of code dependencies for instruction parallelism and the analysis of streaming capabilities, including memory access and caching strategies, as well as parallelization considerations and pipeline analysis. Each architecture is described in all details with its surrounding control structure. As proof of concepts, the architectures are mapped on a Virtex 6 FPGA and results on resource utilization, runtime and image quality are presented and discussed. A special framework allows to scale and port the design to other FPGAs easily and to enable for maximum resource utilization and speedup.The result is streaming architectures that are capable of massive parallelizationwith a minimum in system stalls. It is shown that real-time processing on FPGAs with strict power budgets in time-domain is possible with the GBP (mid-sized images) and the FFBP (any image size with a trade-off in quality), allowing for a UAV scenario.