Azimuth Full-Aperture Processing of Spaceborne Squint SAR Data with Block Varying PRF

Zhuo Zhang, Wei Xu, Pingping Huang, Weixian Tan, Zhiqi Gao, Yaolong QiSensors 2022 · DOI

TL;DR

This paper introduces a novel azimuth pre-processing method for spaceborne squint SAR data that uses a Block Varying Pulse Repetition Frequency (BV-PRF) scheme. This method effectively addresses the challenges of non-uniform sampling and range cell migration inherent in BV-PRF data, enabling full-aperture processing and improved azimuth resolution.

What Problem Does This Paper Address?

Spaceborne Synthetic Aperture Radar (SAR) is crucial for Earth observation, providing high-resolution imagery for applications like ship detection and maritime management. Achieving higher azimuth resolution often involves advanced imaging modes like squint sliding-spotlight SAR. However, these modes introduce significant challenges:

  • Large Range Cell Migration (RCM): The movement of the illuminated area across the radar's view causes targets to shift significantly in range over the acquisition time.
  • Azimuth Signal Non-uniformity: Traditional SAR systems use a fixed Pulse Repetition Frequency (PRF). To handle large RCM and achieve wide swaths, schemes like Continuous PRF (CV-PRF) or Block Varying PRF (BV-PRF) are employed.
  • BV-PRF Specific Issues: While BV-PRF offers a compromise between fixed PRF and CV-PRF, it results in different PRFs for different data blocks. This leads to insufficient and non-uniform sampling within each block, making standard SAR processing difficult.

This paper aims to solve the problem of processing SAR data acquired with the BV-PRF scheme, specifically for squint sliding-spotlight SAR, to achieve full-aperture imaging with uniform azimuth sampling.

Methods

The proposed method focuses on an azimuth pre-processing approach that involves several key steps:

  1. De-skewing and De-ramping: The raw data undergoes range-frequency dependent de-skewing and de-ramping to flatten the distorted 2D spectrum and limit the total Doppler bandwidth within the sampling frequency.
  2. Azimuth Resampling: Each data block, despite having different PRFs, is resampled to achieve a uniform azimuth sampling interval. This is a crucial step to prepare the data for continuous combination.
  3. Azimuth Time Shift: An appropriate azimuth time shift is introduced during resampling to ensure the continuity of the azimuth signal across different blocks.
  4. Azimuth Data Combination: The resampled and time-shifted data blocks are then combined to form a continuous, uniformly sampled azimuth signal.
  5. Doppler History Recovery (Re-skewing): After azimuth combination, a re-skewing operation is performed to recover the original Doppler history, which was altered during the de-skewing process.
  6. Azimuth Mosaic and Doppler Filtering: To address residual aliasing, an azimuth mosaic operation is performed, followed by a range-frequency variant Doppler filter to remove redundant spectrum and obtain the desired 2D spectrum.

The core innovation lies in the resampling of block data with varying PRFs and their subsequent continuous splicing, enabling classical SAR focusing algorithms to process the data effectively.

Datasets

The paper does not explicitly mention using pre-existing datasets. Instead, it relies on simulated SAR raw data generated for point targets to validate the proposed method. The simulation parameters are detailed in Table 1.

Key Insights and Ideas

  • BV-PRF as a Practical Compromise: The BV-PRF scheme is presented as a practical solution that balances the need for wide swaths with the computational complexity of CV-PRF, offering a good trade-off for spaceborne SAR.
  • The Challenge of Non-uniform Sampling: The paper highlights that the varying PRFs in BV-PRF data create significant challenges in azimuth signal processing, necessitating specialized pre-processing techniques.
  • Resampling and Splicing as the Key: The central idea is the intelligent resampling of data blocks with different PRFs and their seamless splicing to create a uniform, continuous azimuth signal. This overcomes the limitations of non-uniform sampling.
  • De-skewing and De-ramping for Spectrum Flattening: The range-frequency dependent de-skewing and de-ramping are critical for preparing the data for uniform sampling by flattening the Doppler spectrum.
  • Efficiency of the Proposed Method: The method is claimed to be more efficient than existing azimuth reconstruction techniques for CV-PRF data, utilizing only complex multiplication and FFT operations.

Key Results

  • Successful Azimuth Pre-processing: The proposed method successfully transforms BV-PRF data into a uniformly sampled azimuth signal, suitable for standard SAR focusing.
  • Improved Imaging Quality: Simulation results demonstrate that the method effectively reconstructs point targets, achieving good resolution and low side-lobe ratios (PSLR and ISLR), as shown in Table 2. For example, the resolution for point targets P1, P2, and P3 are around 2.7-2.8 m in azimuth and 0.94 m in range.
  • Comparison with Other Schemes:
    • Fixed PRF: Fails to obtain complete echo data for targets at the edge of the scene due to large RCM, leading to decreased resolution.
  • CV-PRF: Successfully obtains the whole scene data but has significantly higher computational complexity (dozens of times greater than the proposed method).
  • BV-PRF (Proposed): Achieves successful imaging of the whole scene with significantly lower computational complexity compared to CV-PRF.
  • Validation through Simulations: Figures 16 and 17 showcase the effectiveness of the method in recovering the 2D spectrum and imaging point targets, demonstrating the suppression of aliasing and the successful reconstruction of the full scene.

Limitations and Future Work

  • Sample Number Sensitivity: The paper notes that when the number of samples in the designed BV-PRF scheme is either too small or too large, redundant operations might occur in the proposed azimuth full-aperture processing.
  • Future Research Directions:
    • Designing an "equal and sufficiently small sample number" for the BV-PRF scheme to optimize processing.
  • Investigating the omission of the de-ramping operation to further reduce computational load.
  • Exploring azimuth sub-aperture processing as an alternative strategy for processing BV-PRF data.

Summary updated

BibTeX
@article{zhang2022azimuth,
  author = {Zhuo Zhang and Wei Xu and Pingping Huang and Weixian Tan and Zhiqi Gao and Yaolong Qi},
  title = {{Azimuth Full-Aperture Processing of Spaceborne Squint SAR Data with Block Varying PRF}},
  journal = {Sensors},
  year = {2022},
  doi = {10.3390/s22239328},
}
Figure