1 quotation? According to incomplete statistics, in 20 14, there were more than 2000 urban road collapse accidents in China, and more than 50 cities were affected by road collapse accidents, mainly distributed in more than 20 provinces, autonomous regions and municipalities such as Beijing, Shanghai and Guangdong. According to the materials published by the Ministry of Land and Resources and the Ministry of Water Resources, the urban area affected by road collapse in China is close to 20,000 square kilometers. Road collapse accidents are mainly concentrated in three areas, namely: Yangtze River Delta, Pearl River Delta and North China. The road collapse accident has seriously threatened the public safety of the city and destroyed the normal traffic order. If it can be found and treated in advance, the loss caused by the sudden collapse of the road surface can be minimized. At present, an effective way to quickly detect roads is to find underground diseases of roads by analyzing ground penetrating radar images. ? 1. 1 Purpose and significance of road research Road is the most important infrastructure of a city, and it is also an important channel for personnel exchanges and economic development. With the rapid development of economy and science and technology in China, the mileage of urban roads is increasing, and the traffic is more and more convenient. With the continuous expansion of urban scale and the continuous increase of urban population, ground space can no longer meet people's needs, and underground space has become a useful supplement to ground space. From various pipelines to underground transportation networks, the utilization of underground space tends to be hierarchical and large-scale. Coupled with the diversity and complexity of shallow geological structure, rammed soil under urban roads may be affected by nature and man-made at any time. Therefore, with the rapid development of road construction, road maintenance has also begun to receive attention. 2065438+At the beginning of 2004, the Road Administration Bureau of Beijing Municipal Commission of Communications set up a road maintenance station through the information management software of urban road inspection, which effectively shortened the road disease repair time [1]. The detection and repair of road underground diseases is the key problem of road maintenance. Usually, underground diseases are mainly loose, empty and abnormally rich in water (hereinafter referred to as rich in water). These hidden dangers may lead to diseases such as pumping mud and cracking on the pavement, and serious cavities may even lead to sudden collapse of the pavement. On the morning of September 25th, 20 14, the pavement of the southwest gate of Hua Fu Jiayuan on Huangshandian Road in Beijing collapsed, and half of the houses collapsed and fell into the pit. Fortunately, there were no casualties [2]. The traditional road maintenance and detection methods mainly rely on manual work, which not only has poor accuracy, but also has obvious lag. In recent years, road collapse has occurred from time to time, and people's lives and property have suffered serious losses due to the backward maintenance and testing methods. 20 1 April1day, when Ms. Yang, a citizen of Beijing, passed the service road on the east side of Wuhua Building on Beilishi Road, the pavement suddenly collapsed and fell into a hot water pit. On April 9, Ms. Yang died at the age of 27 because of ineffective treatment [3]. Therefore, it is urgent to use advanced instruments in road detection and realize accurate detection and maintenance by using advanced geophysical technology. Minimize unnecessary losses. Ground penetrating radar is an important part of applied geophysical science. Ground penetrating radar can transmit and receive high-frequency broadband electromagnetic waves in microwave band. Because electromagnetic wave will be reflected at the interface of underground medium, the spatial position of underground target can be obtained by analyzing the waveform characteristics of electromagnetic wave reflected at the interface of underground medium, and the characteristic information such as materials can be formed [4]. ..........？ 1.2 research status and progress at home and abroad The development of ground penetrating radar has gone through 100 years, during which Germans made important contributions. The prototype of ground penetrating radar was born in 1904, and German Hulsemeyer found that electromagnetic waves can detect metal objects on the ground [5]. 19 10 years, Germans Leimbach and L wy first expounded the related technologies of ground penetrating radar and obtained patents. 1926, Huelsen Baker, a German, found that the medium with different dielectric constants would generate electromagnetic wave reflection at its interface, and he put forward the idea of detecting underground targets by using high-frequency electromagnetic wave pulses [6]. During the Second World War (1939-1945), for military purposes and war needs, ground penetrating radar was rapidly developed and applied, and shallow stratum targets were detected. During the Vietnam War (1960), MIT introduced a device for detecting shallow stratum cavities, which was used to find tunnels in Vietnam battlefield [7]. In the same year, CookJ. C made an experiment in the mine with pulse radar, but because the underground medium has stronger electromagnetic wave attenuation characteristics compared with air, and the diversity of geological conditions, the propagation of electromagnetic waves in the underground is much more complicated than that in the air [8]. With the development of electronic information technology, the signal-to-noise ratio of instruments has been greatly improved. The application range of ground penetrating radar has also expanded rapidly, from weak media such as ice and rock salt mines to lossy media such as soil, rock and coal seam. Since 1970s, ground penetrating radar has been applied to the detection of limestone quarries, engineering geology and coal mines. In 1980s, with the rise of the civil market, the carrier-free pulse ground penetrating radar first entered the market, and developed countries competed to take the lead in developing civil ground penetrating radar products. Later, with the continuous upgrading of GPR products, GPR technology has been relatively mature [9]. The application of ground penetrating radar technology in subgrade and pavement detection began in 1980s. 1983, Benson, an American, and others have already carried out relevant research on the settlement and collapse of expressways [10]. In 1984, Rodeick et al. studied the highway cavity detection by using ground penetrating radar [1 1]. 199 1 year, the federal highway administration of the United States has made a series of progress in the application of road engineering, and successfully detected road diseases such as subgrade layered thickness, pavement void and subgrade void. 1993, Japanese M. Sekiguchi and others combined ground penetrating radar with borehole camera to develop a road structure detection system [12]. In 1994, Kim Roddis and others compared the differences of ground penetrating radar data analysis of different types of roads in Kansas 1 1, which are mainly determined by subgrade materials and design structures [13]. 1995, American Laurie industries co., ltd. cooperated with GSSI company to launch the world's first air-coupled high-speed pavement detection radar system within 10 month, and successfully tested it in China, as shown in figure 1.2. .........？ 2 GPR technology and data characteristics? Ground penetrating radar (GPR) is the main method to detect urban road underground diseases at present, which has the advantages of fast detection speed and high accuracy. In this chapter, the wave equation of electromagnetic wave is derived from the electromagnetic field theory. On the basis of theoretical introduction, this paper expounds the principle and present situation of ground penetrating radar technology, and briefly explains the form, characteristics and calibration of ground penetrating radar data. ? 2. 1 electromagnetic field theory 1820, Danish physicist Oster first discovered the effect of current on the magnetic needle, that is, the magnetic effect of current. 1837, the British physicist Faraday proposed for the first time that there are both electric fields and magnetic fields in nature, and both of them can only work within a certain range, turning the elusive "action at a distance" into a "field" that can be understood and studied. Since 1855, British physicist Maxwell has been interested in emerging electromagnetics besides studying elasticity and structural mechanics. He combined his familiar elasticity with electromagnetic phenomena, and expressed the electromagnetic field theory in a concise, symmetrical and perfect mathematical form through three papers, which became the basis of classical electrodynamics in later generations. This is Maxwell equations [55]. On this basis, he predicted the existence of electromagnetic waves in 1865. 1888, German physicist Hertz finally verified the existence of electromagnetic waves through experiments 10 years after Maxwell's death. Classical electrodynamics holds that electrostatic field and static magnetic field are generated by static charge and constant current respectively, which are independent of each other and satisfy their own equations. When the distribution of charge and current changes with time, the electric field and magnetic field are no longer independent of each other, but stimulate and influence each other to form a unified electromagnetic field. Electromagnetic waves are generated from this time-varying electromagnetic field. It can be seen that a group of differential equations derived from Maxwell's equations describing the fluctuation characteristics of electromagnetic fields are called wave equations. Wave equation can describe all kinds of wave phenomena in nature, including shear waves and longitudinal waves, such as sound waves, light waves and water waves. Wave equation is an important mathematical basis for analyzing the propagation of electromagnetic waves in various media. ........？ 2.2 Ground penetrate Radar technology is an electromagnetic instrument used to detect underground dielectric structures. It emits high-frequency broadband (1MHz~ 10GHz) electromagnetic waves through the transmitting antenna, then receives the reflected electromagnetic waves from underground media through the receiving antenna, and finally converts the reflected electromagnetic waves into digital signals through the digital circuit and records them on the storage device. Ground penetrating radar has the advantages of high detection accuracy and high speed, and is an important means of engineering nondestructive testing. At present, GPR manufacturers include Italian Systems Engineering Company (IDS), Swedish Mara Company, Canadian Detector and Software Company (SSI) and American Geophysical Exploration Equipment Company (GSSI), all of which have introduced GPR products for road detection, as shown in Figure 2. 1 Since the 1980s, after more than 30 years of research and development, domestic GPR products have matured and gradually formed their own systems, reaching the world-leading level from signal acquisition to data processing, and enjoying a certain popularity at home and abroad. China University of Mining and Technology (Beijing) State Key Laboratory of Resources and Safe Mining, Chang 'an University Highway College and other units have made important contributions to the theoretical research, instrument development and application promotion of ground penetrating radar. At present, the products entering the market include the urban road detection ground penetrating radar system developed by China University of Mining and Technology (Beijing), as shown in Figure 2.2, and the LTD series ground penetrating radar of the 22nd Research Institute of China Electronics Technology Corporation (Qingdao). Compared with other road nondestructive testing technologies, ground penetrating radar technology has the advantages of fast detection speed and high detection accuracy, so it has become the main means of urban road detection. However, GPR data, like other geophysical detection data, is difficult to interpret, requires high manual interpretation experience and has a long interpretation period, which makes the application and popularization of GPR road detection difficult. In this paper, the ground penetrating radar instrument of China University of Mining and Technology (Beijing) is used to study the algorithm of road underground detection image and underground anomaly identification, which reduces the difficulty of data interpretation and shortens the interpretation period. ............？ 3 physical model design and characteristic measurement of road diseases ... 173. 1 physical model structure ... 173.2 physical model design ... 203.3 characteristic measurement of physical model ...1 Underground cavity detection ... monitoring ... 353.3.3 Monitoring of pavement settlement ... 393.4 This chapter outlines the underground anomaly identification algorithm for 424 urban roads in .......... ... 334. 1 .......... Based on Hilbert marginal spectrum +0. 1 empirical mode decomposition. Hilbert spectrum and marginal spectrum +0.2 experimental results and analysis+0.3+464.2 underground anomaly identification algorithm based on kernel matching pursuit .......... 554.3 This chapter summarizes the underground anomaly measurement algorithm of 675 urban roads, ...................................... radar data preprocessing ... Kloc-0/. 1 Ground Penetrating Radar Data Denoising ... 695. 1.2 Ground Penetrating Radar Data Offset and Seeking ... 695. 1.3 Precise Registration of Ground Penetrating Radar Data ................................ 965.2 Underground Anomaly Measurement Algorithm Based on Periodic Detection ... Application of Urban Road Underground Disease Detection ... 5 urban road underground anomaly measurement algorithm? In the past, the interpretation of urban road underground diseases can only be carried out on one detection result, because the result is often seriously disturbed by the surrounding environment, and the interpretation result has errors. Because the underground cavities that endanger the safety of urban roads will deteriorate with time, it is necessary to detect urban roads many times. By comparing the differences of detection data in different periods, the underground diseases of urban roads are identified. In order to accurately compare the differences of detection data in different periods, it is necessary to accurately measure the underground anomalies of urban roads and determine the location and scope of underground anomalies of urban roads. Specifically, firstly, the noise interference in GPR data is reduced by iterative Myriad filtering denoising algorithm. Then, Shikhov integral migration algorithm is used to shift the signals in the GPR detection image and return them, which effectively improves the accuracy of position and distance calculation. Then, through the fine registration algorithm of GPR images or the standard registration algorithm, the similar areas of the two images correspond to the same position completely. Finally, the appropriate sliding window is selected, and the location and range of underground anomalies are measured by comparing the differences of ground penetrating radar data. ? 5. 1 GPR data preprocessing In the process of GPR image data acquisition, noise interference is an insurmountable phenomenon. With the increase of detection depth, the noise of reflected signal becomes more and more obvious [77-78]. According to the source, noise interference mainly includes the following categories: 1. There is coupling wave interference between transmitting antenna and receiving antenna. Even if shielding materials such as metal are used, there is still no guarantee that the electromagnetic wave of the transmitting antenna will not be coupled to the receiving antenna; Second, the impedance of the transmitting antenna and the transmitting cable do not match. Impedance matching must be considered when the transmitting antenna is connected to the transmitting cable, otherwise it will lead to energy loss and form standing wave interference signal; Third, the oscillation interference between the signal emitted by the antenna and the antenna shield. For broadband antennas, it is difficult for the shielding body to ensure good shielding of all frequency signals, and there is often oscillation interference between the antenna transmitting signals and the antenna shielding body; Fourth, the antenna feed point reflects signal interference. The feed point is the connection point between the antenna and the feed line. Although some reflected signals can be absorbed by absorbing materials, some signals will still cause standing wave interference. Fifth, the sidelobe interference of the transmitted pulse signal. Theoretically, there are no side lobes in the transmitted pulse signal. In reality, it is impossible to have only main lobe signals, and these sidelobe signals can also cause interference.
........？ Conclusion? In this paper, the ground penetrating radar (GPR) detection image is taken as the research object, and the related technical difficulties in the application of GPR in the detection of urban road underground diseases are emphatically analyzed, and the difficulties of GPR image interpretation, such as great difficulty, high requirement for manual interpretation experience and long interpretation period, are emphatically broken. Focusing on the goal of underground anomaly identification and measurement of urban roads, the dynamic evolution model of underground cavities in urban roads is established, and some key problems such as anomaly identification and measurement based on ground penetrating radar images are studied. The main work of this paper can be summarized as follows: 1. Through the physical model experiment of urban road underground diseases, the following conclusions can be drawn: when underground construction and other disturbances occur, on the one hand, underground cavities are formed due to the disturbance, and the surrounding soil is unevenly stressed, which leads to the decrease of density, thus causing pavement settlement. On the other hand, the formation of underground cavities will lead to the contact between underground soil and air, and the continuous evaporation of water will lead to the decrease of compactness, which in turn will lead to pavement settlement. 2. Through the research on the underground anomaly identification algorithm of urban roads, the following conclusions can be drawn: 1. Because both underground cavities and metal pipes will cause the change of Hilbert marginal spectrum, the underground anomaly identification algorithm based on Hilbert marginal spectrum can be used not only for the detection of underground cavities, but also for the detection of metal pipes. The underground anomaly identification algorithm based on Hilbert marginal spectrum can estimate the density of a single sandy silt model through the amplitude of marginal spectrum, and then find underground anomalies. In the process of urban road underground detection, due to the influence of underground pipelines and structures, the density estimated by the above algorithm may have errors. 2. Underground anomaly identification algorithm based on kernel matching pursuit. The density is estimated by the proportion of wavelet kernel function, so as to find underground anomalies. The estimated results of average density will not be disturbed by metal pipes, and it has a good application prospect in detecting underground porosity and cavity diseases. 3. Through the research on the measurement algorithm of urban underground anomalies, we can get the following conclusions: 1. Through iterative Myriad filtering denoising algorithm, the noise interference in ground penetrating radar data is reduced, and the optimal signal-to-noise ratio is 28.357dB, which is 3.5dB higher than that of Myriad filtering denoising algorithm. Therefore, compared with Myriad filtering denoising algorithm, iterative Myriad filtering denoising algorithm can achieve better filtering effect. 2. Through Shikhov integral migration algorithm, the signal in the GPR detection image can be migrated, and the migration effect is the best when the parameter is 30. 3. Maintain data consistency through fine registration algorithm of GPR data or inter-standard registration algorithm. Experiments show that the correlation coefficient between the recovered GPR data and the original data can still reach above 0.9 when the track is lost by 90% through fine registration and standard room registration. This can partially eliminate the consistency differences caused by data loss, acquisition software settings, water content changes and other factors. Because registration is achieved by the difference between horizontal and vertical directions, the damage to signal characteristics is reduced. Reference .......... (omitted)