3D Ground-Penetrating Radar Void Detection Method Based on Spatial Convergence

Abstract

In order to solve the problem of void feature detection of 3D ground-penetrating radar data, the test of void detection in polystyrene cubic simulation road is carried out in this paper. The B-scanning image was processed by binarization, corrosion, expansion and thin length, and the suspected void region was obtained, According to the characteristics of multi-channel detection of 3D ground-penetrating radar, a method is proposed to divide the B-scanning image of suspected void into several clusters and calculate the convergence of each cluster to identify and determine the void signal. The detection method has been verified in the engineering, The results show that in the process of the field test and verification, the algorithm of 3D ground-penetrating radar data void characteristic detection accurately recognizes 4 sites with void potential. Also, the 3D matching test successfully eliminated 8 suspected void regions, indicating that the 3D ground-penetrating void feature detection algorithm has excellent robustness in the task of void segmentation of radar images. The technique can detect the voids inside city roads accurately and fast which have great significance in reducing casualties and property damage.