Research On Subgrade Detection Technology Based On Ground Penetrating Radar

In recent years,in order to adapt to economic development and social needs,the state has increased its investment in road construction.The development of road construction directly affects the development of the national economy.The quality detection of road subgrade is the most important part of road construction,and the results of subgrade inspection have a guiding role in the construction and maintenance of roads.Whether in China or abroad,the ground penetrating radar method with high efficiency,continuous and non-destructive advantages has been widely used in subgrade quality inspection.However,the interpretation technology of the subgrade radar image in the later stage of detection is still not perfect enough.This thesis studies the subgrade detection technology based on the ground penetrating radar method,in order to supplement the interpretation algorithm of the subgrade radar image in the later stage of detection,and proposes an effective location technology of subgrade disease area.Based on the characteristics of subgrade structure and the basic characteristics of subgrade disease,this thesis firstly simulates the subgrade radar echo data,and then studies the data processing algorithm of subgrade radar echo using the simulation data.This thesis proposes a clutter suppression method combining principal component analysis and mean filter to preprocess the echo image,and proposes an improved spectral residual method to accurately locate the subgrade disease area,the Markov chain theory is introduced to realize the layer information tracking of the subgrade.Finally,the data processing software of subgrade detection echo is developed.In the aspect of image preprocessing,the characteristics of clutter contained in the echo signal of subgrade radar are analyzed.The clutters such as direct wave,coupled wave and noise are classified into two categories according to the characteristics of its energy.An improved principal component analysis method is proposed to preprocess the radar echo image,which effectively suppresses the clutter and noise,greatly improves the image quality of the subgrade radar echo,and provides favorable conditions for the subgrade disease area location and layer information tracking.In the aspect of subgrade disease area localization,a new spectral residual algorithm based on multi-resolution monogenic signal analysis is proposed to extract the region of interest in the radar image and improve the location accuracy of the subgrade disease area.The algorithm comprehensively considers the characteristics of the radar echo signals in the subgrade disease area,extracts the amplitude component of the subgrade radar image by multi-resolution monogenic signal analysis algorithm,and then processes the radar amplitude component map by using the spectral residual method in the field of visual saliency detection.This method does not need prior knowledge,and can accurately and quickly extract the region of interest of the subgrade radar echo image,and locate the subgrade disease area.In the aspect of subgrade layer tracking,the Markov chain theory of subgrade radar echo time delay is studied.Based on this theory,the detection and tracking of subgrade layer information is realized.The simulation and measured data results show that the algorithm used in this paper has good processing results and accurate results.Finally,after completing the research of subgrade radar data processing algorithm,this thesis developed a processing software of subgrade echo data based on ground penetrating radar,which realized the common data processing method of subgrade radar signal and the algorithm proposed by this thesis.