The underground runoff caused by precipitation in the mountain area erodes the fragile soil structure,which leads to the uneven settlement of the base of the high-voltage tower in this area,which poses a serious hidden danger to the safety of the transmission line.We use ground penetrating radar(GPR)technology to detect underground areas with high water cut and cavities.However,the traditional GPR technology is vulnerable to clutter,resulting in low positioning accuracy.To solve this problem,a clutter suppression algorithm based on frequency domain entropy detection is proposed.The algorithm measures the degree of confusion of the system by entropy.The cavities and high-water-content areas in the soil stratification will make the echo signal energy concentrated,the orderliness improved,and the entropy value reduced.Based on the finite-difference time-domain(FDTD)method and gprMax,this paper simulates the cavity and high water-cut area of soil layer.Comparing the F-K filtering and entropy detection methods,the results show that the frequency domain entropy test can better realize the focusing of the echo energy,suppress the clutter,discriminate the interference and target,and effectively improve the detection performance of the underground layered discontinuous structure and high water content area.