E-GEOS (IT)
The combined use of traditional satellite SAR and Inverse SAR (ISAR) techniques applied to SAR data has been investigated to assess their impact on feature extraction processes and improvements in data fusion results. Unlike SAR, which uses the radar platform’s motion to generate a synthetic aperture for imaging, ISAR relies on the motion of the target itself to create the synthetic aperture antenna. ISAR images of moving targets, processed as though the radar platform is stationary, are well-focused, contrasting with standard SAR images of moving targets, where motion remains uncompensated, leading to blurring. Additionally, ISAR enables the retrieval of target motion parameters such as speed and direction, offering valuable insights into target characterization and activities.
The project, led by e-GEOS(IT) with CNIT (IT) and Aresys (IT) was successfully finalized in December 2024. As part of ESA’s research and development activities on advanced processing, performance assessments have addressed the enhancement of SAR image contrast and the accuracy of target velocity estimation.
In the project, innovative techniques were explored and tested across various use cases, including the application of Back Projection (BP) as an alternative to traditional Image Contrast Based Algorithms (ICBA), and the integration of autofocus algorithms, with particular focus on the Ash autofocus method. A fully polarimetric-based algorithm was preliminarily designed to enhance the detection of moving targets within clutter by utilizing polarimetry as an additional domain, beyond spatial domains like range and cross-range, to improve target-clutter distinction. Additionally, a convolutional neural network (CNN) was designed to process complex ISAR data and extract significant patterns related to target motion. Synthetic data were generated to train and evaluate the AI model. Several use cases were implemented to test the service chains in different scenarios, addressing both maritime awareness and land surveillance. Promising results have also been achieved for land scenarios, focusing on moving targets such as trains and cranes, with significant improvements in image contrast and sharpness.
The research on combined SAR/ISAR service chains aims to enhance the interpretability of SAR data by enabling detailed target extraction, reducing false alarms, and improving classification and recognition capabilities to support maritime surveillance and awareness. The integration of SAR/ISAR techniques supports the quantification of target velocity from SAR images, complementing information provided by AIS data, while also enabling the extraction of information when no AIS data is available. ISAR-based refocusing is fundamental for enabling automatic classification and recognition of moving targets in SAR imagery.
The ISAR techniques studied in the project can also enhance the application of advanced signal processing methods, such as micro-Doppler estimation and video SAR formation, further expanding their potential applications in both maritime and land-based scenarios.
