研究目的
To address the problem of saturation in SAR raw data caused by finite quantization bits and unpredictable scene scattering characteristics, which leads to non-linear distortion, false targets, and degraded SNR, by proposing a recovery method based on the non-linear characteristic of compressed sensing to restore saturated parts to unsaturation while preserving unsaturated parts.
研究成果
The recovery method based on compressed sensing effectively depresses saturation error in SAR raw data, improving image quality and radiometric accuracy. Simulations show reduced relative error and elimination of false targets, particularly in coastal areas. Future work could involve automating parameter tuning and extending the method to non-sparse scenes.
研究不足
The proposed method assumes scene sparsity, which may not hold for all scenarios. The termination condition in the iterative algorithm requires manual tuning of the epsilon parameter for each application, and there is no automated way to set it currently. Additionally, the method is specifically validated for coastal areas where saturation is common, limiting its general applicability to other terrains.
1：Experimental Design and Method Selection:
The methodology involves using compressed sensing (CS) with l1-norm optimization to recover saturated SAR raw data. It leverages the non-linear characteristic of CS for reconstruction at Nyquist sampling rates, focusing on range dimension signals.
2：Sample Selection and Data Sources:
Simulations are conducted using synthetic data for one point target, multiple point targets, and a coastal area scene based on TerraSAR sensor parameters. Saturation rates are varied to test the method.
3：List of Experimental Equipment and Materials:
No specific physical equipment is mentioned; the work is simulation-based using computational tools for signal processing and optimization algorithms.
4：Experimental Procedures and Operational Workflow:
The process includes constructing a signal model for range dimension, formulating an l1-norm optimization problem, applying an iterative algorithm to solve it with initial and termination conditions, and then using chirp scaling for imaging. Saturation is simulated by setting threshold values based on input signal RMS.
5：Data Analysis Methods:
Relative error (Er) is calculated to evaluate reconstruction accuracy by comparing results from unsaturated and recovered saturated data. Statistical analysis involves plotting Er against saturation rates.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
