Investigation of water pipeline leakage is a matter of great significance for water resources management. Among various approaches, ground-penetrating radar (GPR) attracts increasing attention as a fast and noninvasive technique. However, this method suffers from uncertainties in the stage of interpretation because diffractions from the leak region often are too weak to be identified. To address this problem, we have developed the use of diffraction imaging to enhance imaging and characterization of water leakage. Numerical seepage simulation allows us to establish an equivalent computational analog for a known leak condition. After the collection of parallel GPR survey lines, we apply diffraction separation and imaging techniques on 2D and 3D data. Although the actual situation is generally well reproduced by 2D and 3D results, 3D diffraction imaging provides improved mapping of leakage compared with its 2D counterpart. We then apply the method to field data collected at a training base for water pipeline leakage control in Shaoxing, China. The prediction from imaging results corresponds well with the true position and type of leakage. The laboratory and field experiments substantiate the viability of GPR diffraction imaging and illustrate the potential of using GPR to detect small water leaks.

SUMMARY: We present a novel strategy for performing joint inversion with guided fuzzy c-means (GFCM) clustering coupling and apply it to electrical resistivity tomography (ERT) and ambient noise surface wave (ANSW) data. To accurately extract a priori clustering information, we use density peak clustering (DPC) rather than fuzzy c-means (FCM). The number and centres of resistivity and shear-wave velocity a priori clusters are extracted by DPC and then used to guide the joint inversion with the GFCM clustering coupling of ERT and ANSW data. Synthetic and field data are used to evaluate the flow and algorithm of DPC-GFCM clustering joint inversion. The results of synthetic examples show that the models recovered by the DPC-GFCM clustering joint inversion are nearly the same as the true models and are more accurate than those inverted using individual inversion and FCM-GFCM clustering joint inversion. In the field case, the depths of the stratigraphic interfaces shown in the resistivity and shear-wave velocity models inverted by DPC-GFCM clustering joint inversion are nearly consistent with those from the drilling data. In contrast, the strata recovered by the individual inversion and FCM-GFCM clustering joint inversion significantly differ from the drilling results. Both the synthetic and field examples verify the effectiveness of the DPC-GFCM clustering coupling method used for the joint inversion of ERT and ANSW data acquired from the near surface with strong heterogeneity. This novel approach can also be applied to other types of geophysical data.

Road collapses seriously threaten human safety in cities. The effective and rapid identification of hidden urban road hazards using 3D ground-penetrating radar (3D-GPR) data is crucial for preventing and controlling road collapse accidents. We propose a novel intelligent identification method for urban road hazards using a backpropagation (BP) artificial neural network based on multiscale gray-level co-occurrence matrix (GLCM) attributes derived from 3D-GPR data. The characteristic differences in multiscale GLCM attributes between hidden hazards and other detected targets are leveraged to construct a road hazard identification model using a BP neural network. This model is applied to identify hazards from 3D-GPR data, enabling rapid localization of hidden urban road hazards. Initially, GPR data are processed using zero-time correction, background suppression, and filtering. Next, labeled data are created using the radar responses of known hazards, wells, pipelines, and strata. Subsequently, after applying gradient enhancement processing to the GPR data, the GLCM attributes are calculated at different scales to distinguish hazards in three-dimensional space from other target bodies. Finally, a neural network model for hidden hazard identification is constructed using a backpropagation (BP) artificial neural network and applied to identify hidden hazards beneath urban roads from GPR measurement data. The results show that the prediction accuracy exceeds 85% using the proposed approach. The combination of multi-scale GLCM attributes and BP neural network is valuable and effective for the detection and identification of hidden hazards beneath urban roads.