| Along with the rapid development of the expressway in our country,detecting the quality of new roads is increasingly arduous day by day。During the course of maintaining pavements,workload of quality evaluation of pavements is increasing sharply。The traditional detection method is slow。Points for detecting are at random and the density is sparse。It's difficult to find the recessive distress existing in pavements accurately。The traditional detection destroys the pavement surface。It is apt to damage the pavement。As a nondestructive detecting equipment,GPR has many features,such as fast,continuous,high-precision,high-resolution and real-time formation of image。On the base of detecting thickness of the pavement quickly and accurately,GPR also can detect distress of pavements successfully and is favorable to maintain pavements。It can offer a nondestructive detecting method high-efficiently for traffic department of pavements。In the practical application of GPR,the strong surface reflection wave,the direct coupling wave,the noise disturbance and other factors may cover up the useful information,which will seriously affect quality of the image。The image information acquired need the expert to explain,which makes it a very subjective interpretation。People can not distinguish the types of distresss of pavements,so it is unable to take corresponding emergency measures。In order to abstract the useful information accurately from the obtained GPR materials,detect and identify recessive distress of pavements accurately,signal processing and target identification of GPR materials are necessary and have great theoretical and practical significance。In recent years,signal processing and target identification technology of Ground Penetrating Radar have made tremendous achievements。Based on previous experience , combining the high-tech projects of JiLin Provincial Science&Technology Department --"Intelligent Integrated Detection Vehicle of Subgrade and Pavement",this paper carried on a series of research against the technical difficulties of GPR signal processing and target identification methods。Our research work is as follows:(1) Introduced the wide application and research condition of Ground Penetrating Radar at home and abroad,and summarized the current situation of the Ground Penetrating Radar signal processing。(2) In view of signal disturbance of Ground Penetrating Radar,we use the wavelet transform to de-noise signals of Ground Penetrating Radar in order to improve the quality of GPR images。It makes early preparations of the geological interpretation accurately and reasonably。(3) Summarized and compared the method of estimating the wave speed from previous,chose the method of reflection coefficient as the first-selected method for estimating the thickness of pavement。We carried on an objective appraisal of the detecting accuracy of Ground Penetrating Radar from theory and experiments。The experiments show that the detecting accuracy of the Ground Penetrating Radar is in the range of 5%,which is satisfied with the requirements of detecting the thickness of pavements。(4) In view of the flaw of Pavement,we first analysis of characteristics of Pavement flaw,extract the wavelet energy spectrum and welch power spectrum of Ground Penetrating Radar signal as characteristics,and use wavelet neural networks and learning vector quantization for the flaw of pavement to carry on the target identification。The experiments show that learning vector quantization neural network have good function of classification, can identify the flaw of pavement。(5) Through detecting the practical pavement,we test the global capability of the intelligence detecting vehicle。Focusing on describing GPR detection process and interpreting the detecting image。(6) According to the practical need of the subject,we designed a series of typical models in order to simulate the distress of the pavement, such as voids ect。we detect the models and obtain good results。...
|
PDF Full Text Request