Design of Large-Scale Spaceborne SAR Processing System Based on Multi-Node Collaboration

The onboard processing technology of spaceborne Synthetic Aperture Radar （SAR） is essential for enhancing the efficiency of remote sensing applications. It provides a critical solution to the bottleneck in space-to-ground data transmission bandwidth and addresses delays associated with data processing in traditional models. This technology plays a vital role in various domains， particularly in military reconnaissance， where it enables the acquisition of high-resolution imagery for intelligence gathering， and in disaster emergency monitoring， where it allows for the timely detection and assessment of disasters. However， current onboard processing hardware systems exhibit significant limitations in terms of functional coverage and their ability to handle complex tasks. Most existing research focuses on individual processing steps or relatively simple tasks， making it difficult to adequately meet the demands of modern spaceborne missions. These missions often require the integration of multiple functions and high-performance processing capabilities that current systems cannot provide. To overcome these limitations， this paper presents the design of a large-scale onboard processing system for spaceborne SAR. The proposed system supports a variety of operations， including data preprocessing， SAR imaging， ship detection， refocusing， target recognition， and geometric correction. Of particular note are two advanced capabilities： wide-area ship detection and recognition. A multi-task compatible approach based on multi-node collaborative reconstruction underpins the system’s design. This method informs the development of a detailed hardware architecture comprising several dedicated processing boards. The interactive master control board manages data input， preprocessing， and distribution， serving as the central hub for system-wide data flow. The detection board is responsible for target detection and refocusing tasks， aiming to accurately identify and re-image blurred moving targets. The imaging board， equipped with dedicated System-on-Chip （SoC） chips， performs high-speed SAR imaging to ensure the production of high-quality， high-resolution images. The recognition board handles false alarm rejection and target recognition， providing classification information for detected objects. Specialized data flow and pipeline designs have been developed to support wide-area ship detection and recognition functions. In the wide-area ship detection process， preprocessed data is distributed to the appropriate boards， followed by SAR imaging， target detection， false alarm rejection， moving target localization， and geometric correction. For wide-area ship recognition， the process includes additional steps， refocusing and recognition， after detection to enable precise identification. The system’s functionality and performance were verified through a series of experiments using a simulated signal source to generate spaceborne SAR echo data. The experimental results demonstrate that the SAR imaging function meets the required standards in terms of resolution， peak sidelobe ratio （PSLR）， and integrated sidelobe ratio （ISLR）. The system achieves a target detection rate of 93.7%， with a false alarm density of 2.96 per 10000 square kilometers. In terms of processing efficiency， the wide-area ship detection function reaches a 1∶2 near-real-time processing level， whereas the wide-area ship recognition function achieves a 1∶2.5 near-real-time level. In conclusion， the proposed system exhibits excellent processing quality and impressive imaging speed， offering a valuable reference for the future development of large-scale onboard processing technologies for spaceborne SAR applications.