Pipeline structure and leakage detection method of full polarimetric ground penetrating radar based on model and eigenvalue decomposition

Leaks in water distribution systems occur frequently and are difficult to detect, resulting in significant economic losses and resource wastage. Traditional ground-penetrating radar (GPR) requires complex data processing procedures and expert interpretation to identify weak scattering signals from leakage zones and lacks a standardized interpretation framework, limiting the development of automated leakage fault detection. This paper proposes a pipeline structure and leakage detection (PSLD) approach using full-polarimetric ground penetrating radar (FP-GPR) based on model and eigenvalue decomposition. The improved polarization decomposition technique is utilized to assess pipelines' structural changes and leakage faults, and a robust planar classifier is constructed. Numerical simulations and model experiments validate the effectiveness of the proposed method. The PSLD technique demonstrates the potential for automated inspection of complex pipeline systems.