Based On The Knowledge Of The Active Faults Auxiliary Recognition

Active fault is a disaster in the world within the general development of disasters, its frequency and the damage caused by the disaster and the damage caused by the degree of layer is gradually intensified. However, the western region of modern crustal activities strong, very active faults development, and its strength is far greater than the number of the eastern region, Highway construction projects will face a serious threat of active fault disasters. And to identify and study fault rupture is the basis and premise, Fault current activities of the entire process to discriminate manual, and a small personal experience, efficiency, relatively underground, and discriminate accuracy and discriminate methods and personal experience is more related. And remote sensing image is reflected in the images from the mainland region of the electromagnetic radiation energy, a clear physical meaning. Remote sensing image data pixel brightness value of the changes in size and features are caused by the main types of changes. It makes the spectral information of remote sensing to identify target objects Based on the characteristics as possible. Through the decision tree in the different nodes of different remote sensing data and appropriate methods of data analysis can be clear, different levels to gradually put the study of distinction, identify.Based Ningxia Haiyuan active fault zone study area as an example, the focus on knowledge-based image classification start the process of research, Discussion on the use of decision tree method for the identification of active faults in the procedure and method in the study area of active faults in the imaging features were analyzed, and the relevant characteristics of the analysis and expression; Through a lot of feature extraction and selection of research and experiment, the traditional method comparison and analysis Mining in a suitable test area model of decision tree classification to characteristics, methods and parameters; Finally, the decision tree classification of experimental and traditional classification methods are compared.According to the final decision tree generated in the study area for information extraction, to highlight good surface boundaries of the two different colors line extension highlighted the good borders, combine field observation point completed the extraction of the active faults, to achieve better results. Note knowledge-based information extraction methods in fault information extraction activities have certain advantages. In summary, the study of knowledge-based remote sensing images active fault extraction method is feasible.