Visualization Of Data Texture Of Ground Penetrating Radar

With the development of economy, the demand of human society for oil, coal, natural gasand other resources is growing faster, and these energy substances are hidden in undergroundreservoirs that people cannot be obtained directly. GPR (Ground Penetrating Radar, GPR) is aneffective means of detecting a subsurface targets in recent years, using GPR collected data foranalysis which we can understand the basic information of subsurface and reservoir. But thespread of radar is very complex underground, many kinds of noise and interference becomesvery serious underground while we use transmitting antenna. Therefore, to identify the extrackvaluable information correctly is essential for personnel in all kinds of clutter and noise, and italso an important part of GPR interpretation.Data visualization technology (Data Visualization Technology) is a visual expression ofthe data interaction to enhance cognitive technology. It can made the data of not invisible ordifficult to display directly mopping for perceptible symbols, colors, graphics, textures etc,then to enhance the efficiency of the identification data and transmission useful data efficiently.The data is the processing target for data visualization. According to the data of processing, datavisualization consists of two branches of scientific visualization and information visualization.GPR data is abstract data of unstructured, non-geometric data, it belonging to the category ofinformation visualization. Therefore this research is proposed for information visualization ofGPR data.In this paper, the GPR data visualization of texture techniques are discussed, and thenimproved an encoding method for texture feature extract based on LBP (Local Binary Patterns,LBP), the final we will be improved LBP used in GPR data. Following, this paper proved anGLCM (Gray Level Cooccurrence Matrix) method for GPR data of texture features,and thencombined with PCA(Principal Component Analysis) for FCM clustering, next the human canget the information from images direct. Through these two experiments which we can introductand description the GPR data visualization of texture. In this paper, the main work includes thefollowing aspects:1. According to the characteristics of the ground penetrating radar data, this paper proposesan improved LBP coding method for the ground penetrating radar data to extract texturefeaturesIn this paper, we analysis ground-penetrating radar data deeply, find the GPR data has ahigh similarity in adjacent data set, but the same physical medium such as sand, stones,minerals has the same structure below the surface. We proposed the improvement of LBPencoding GPR data based on the original LBP encoding method, and then comparing the test results show that the improved LBP encoding to achieve a better test result.2. Made the method of GLCM and PCA applied to the the GPR data visualization oftextureIn this paper, the method of GLCM texture extraction is applied to GPR data visualization,We can extracte multiple features data based on the GLCM. Although, these features may wellreflect some of the texture, but the number of high-dimension feature set is much more, it didnot get a good visual effect, and it take a great resource and takes up much unnecessary time.Therefore, this paper proposes a method of combining the GLCM and PCA for GPR datavisualization of tectyre. After experimental analysis, the method of combine GLCM and PCA isbetter than GLCM approach.