研究目的
To develop and apply a methodology based on the GUM for analyzing and quantifying uncertainties in SAR remote sensing data processing, enabling traceability to international standards.
研究成果
The methodology successfully quantifies uncertainties in SAR data processing for Level-0 and Level-1, providing uncertainty budgets and enabling traceability to SI units. It represents a novelty in the SAR field and can be extended to more complex scenarios and additional uncertainty sources in future work.
研究不足
The study is limited to a simple simulated scene with one target; more complex scenes are not addressed. Only additive noise is considered as a variable input in this initial application; other input parameters are under investigation. The methodology is applied to stripmap acquisition mode and may not generalize to other SAR modes.
1：Experimental Design and Method Selection:
The methodology involves applying the GUM framework to SAR data processing, using Monte Carlo method (MCM) for uncertainty propagation.
2：Sample Selection and Data Sources:
A simulated SAR scene with one pixel target and absorbing background is used, based on ERS-2 ancillary data.
3：List of Experimental Equipment and Materials:
MATLAB software for implementing algorithms to process simulated RAW and SLC data.
4：Experimental Procedures and Operational Workflow:
Implement mathematical models for Level-0 and Level-1 SAR data in MATLAB, perform random draws from input parameter PDFs (e.g., Gaussian noise), run simulations iteratively (e.g., 50,000 trials), and aggregate outputs to define distribution functions and PDFs for pixel magnitude.
5：Data Analysis Methods:
Use statistical analysis to estimate mean, standard deviation (uncertainty), and coverage intervals for the output quantities from the PDFs.
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
