A General Framework for CFAR Detection in PolSAR Imagery Based on Quadratic Statistics

In the field of target detection in polarimetric synthetic aperture Radar (PolSAR) imagery, the constant false alarm rate (CFAR) algorithm is renowned for its operability and high interpretability. Given the challenges faced by deep learning methods in scenarios with limited labeled data and insufficient prior information, CFAR detection techniques remain vital in resource-constrained environments requiring rapid response. However, in the context of CFAR detection in PolSAR imagery, traditional intensity-based statistical modeling approaches, such as gamma distribution, generalized gamma distribution, and log-normal distribution, become inadequate when handling feature maps generated by nonpositive definite transformation matrices. In addition, some nonparametric methods suffer from a lack of robust statistical theoretical support, resulting in insufficient robustness. To effectively address this core issue, this study proposes a general framework for CFAR based on quadratic statistics, employing both the numerical solution of the Gil-Pelaez inversion formula and Monte Carlo reinforcement as strategies to determine detection thresholds. Experimental results from both simulated and real data demonstrate that this approach exhibits superior detection performance across diverse application scenarios, significantly improving the accuracy and robustness of target detection in PolSAR images. The key strength of this framework lies in providing a novel, unified solution for PolSAR image processing, effectively integrating the rigor of statistical analysis for practical applications.