Mapping the Buried Cable by Ground Penetrating Radar and Gaussian-Process Regression

With the rapid expansion of urban areas and increasingly use electricity, need for locating buried cables is becoming urgent. In this paper, a noval method to locate underground based on Ground Penetrating Radar (GPR) Gaussian-process regression proposed. Firstly, coordinate system detected area conducted, input output are determined. The GPR moved along established parallel detection lines, hyperbolic signatures generated by identified fitted, thus positions depths some points cable could be derived. On basis derived cable, clustering fitting algorithm proposed find most likely locations cables. Furthermore, confidence intervals cable's also obtained. Both position depth noises taken into account in our method, ensuring robustness feasibility different environments equipments. Experiments real-world datasets obtained results demonstrate effectiveness method.