Research And Implementation On Stack Velocity Analysis And Singularity Detection

This thesis is a part of "Research on Signal Processing and Video Methods for Phased-array Ground Penetrating Radar (PGPR) (Project Number: 2001AA132050-03)", which is a sub-project of the Nation 863 Project "Exploring Technology of High Resolving Power Ability for Penetrating SurfaceLayer----Phased-array Ground Penetrating Radar" (Project Number:2001AA132050). It is difficult to solve the contradiction between exploring depth and resolution for the actual Ground Penetrating Radar. PGPR can introduce phased array technology into traditional GPR area to resolve the problems. Now PGPR is in a phase of development, and relative theory, information and reference data are very hard to find, so how to analyze and explain the PGPR echo signals becomes very difficult.Velocity is a very important parameter during the penetrating process. It can provide parameters for superposition processing of multi-channel data, migration and layer-velocity. The singularities of Ground Penetrating Radar (GPR) echo wave signals can show the characters of objects or the layer structures under the ground sufficiently. This thesis is about the research of stack velocity analysis and singularity detection.Stack velocity analysis needs to use the analysis and processing method of earthquake signal. This thesis makes use of especial structure and working style of PGPR antenna-array and brings forward an advanced stack velocity analysis. This method is also combined with image enhancement technology, including linear switching and non-linear switching. The advanced method can make velocity spectral image look better. In order to test the efficiency of algorithm, this thesis firstly uses the simulated data by software, and then uses the road test data for the first time of the Three Gorges. This method can get the velocity spectral image and we can get some important information from velocity spectral image, such as stack velocity and the situation of underground strong echo .We also can describe underground configuration from velocity spectral image, and the velocity spectral image provides some important parameters for superposition processing.The singularity detection method of PGPR echo wave signals is based onMallat's singularity detection method based on wavelet maxima, we compare the differences of wavelet maxima in multi-scale decomposition between signal and noise, depending on this difference we can restrain part of the noise disturbance . We use abundant echo data and analyze the relationship of the PGPR echo wave neighboring scan signals and find that the neighboring scan signals are similar. We also find the relationship between weak and strong or early and late echo signals with arrangement of antenna cells. Based on all relationships we get an advanced method, this improved algorithm can restrain the noises' disturbances more efficiently. We also can get the exact positions of the singularities and bring important foundation for analyzing underground configuration.