Nondestructive Detection and Identification Method of Highway Structure Layer Based on Multi-Dimensional Scale Feature Enhancement

Aiming at the problem of weak target leakage caused by the detection echo of the ground penetrating radar（GPR） interfered by the complex underground environment, a nondestructive detection and identification method for highway structure layer based on multi-dimensional scale feature enhancement was proposed. In the method, the texture images of different target echoes in the ground penetrating radar echo to the frequency and direction scale space were matched, and the characteristic of different target signals by adjusting the different scale parameters of the two-dimensional Gabor filter kernel function was enhanced. The weak targets in the ground penetrating radar echo by using different frequency and direction scale features were extracted. The effectiveness of the proposed method is verified by two sets of measured data, and the missed detection rate of weak targets is reduced by 25%.