Ground-Penetrating Radar (GPR)-Based Mapping of Underground Pipeline Network Distribution

As the “lifeline” of urban operation, buried pipelines play an essential role in modern society. Due to the lack of comprehensive pipeline information, pipeline damage accidents often occur in urban construction. This article proposes a ground-penetrating radar (GPR)-based mapping method, including image processing and target detection technique, which applies to underground pipeline networks. The main process of this method is to pre-process the echo data by reasonably arranging the measurement line, fit the hyperbola to extract the peak vertex, and finally fit multiple detection points to realize the pipeline distribution fitting. An innovative fitting method based on random sample consensus (RANSAC) and the
$K$
-means++ algorithm was proposed for GPR testing. The buried depth was estimated and combined with the fitting results to map the underground pipeline network distribution. The effectiveness of the proposed innovative pipeline fitting method was verified through numerical simulation. Then, in situ experiments were conducted on single and multiple pipelines. The numerical and experimental results show that the proposed method can accurately locate buried pipes and effectively reconstruct the direction distribution of underground pipeline networks.