Study On Extraction Of Fault Structure By Multi - Source Remote Sensing Image

Many multi-stage faults of different scales and characters were developed in Qinling orogenic belt,during which the regional trunk faults were well researched.However,the research of associated secondary faults had relatively become the weak point and their characters,occurrence and relations with regional trunk faults were not yet clear.Danfeng,Qinling was selected as the study area because of its landforms, mainly controlled by mountains, plains and well developed faults,which could be easy for the implement and validation of remote sensing extraction methods.In addition,high vegetation coverage rate, another notable feature of study area, could offer a positive reference for the faults extraction of same conditions.The characters of rich spectral information, high spatial resolution and wide field view of remote sensing data could make the faults,which were presented in a form of linear information on remote sensing images,easy to recognize.Compared to the conventional field work,remote sensing extraction methods will not be limited by climate, topography and other factors.Meanwhile,these methods could be taken as the supplementary research techniques owning to the macroscopic characteristic of remote sensing images.Combined with the field investigation,the distribution and characters of the faults were revealed through the extraction methods of multi-source remote sensing images,such as IRS-P6, ETM+, OLI, Google Earth and SRTM.The following methods were adopted for the interpretation and extraction of the faults.Firstly,color synthesis, principal component analysis, directional filtering, image fusion and ratio calculation, combined with three-dimensional display, were adopted to enhance the spatial structure information and spectral information of faults on multi-source remote sensing images and establish the interpretation signs of study area.Based on this,interpretation results were tested on fields. Secondly,automatic extraction method of median filter-edge detection-Hough transform was directly utilized to detect the faults of study area.The results suggested a good effect of these methods.Major faults in study area were commendably extracted through the method of automatic extraction.Moreover,the properties of these faults were judged on the basis of interpretation signs of synchronous bending of river systems,ridge twist.etc.The distribution characteristics and properties of major faults,such as Meijiawuchang-Chenjiawuchang Fault(F2),Wafangcun-Anjiping Fault(F3),Huiyu-Damiaogou Fault(F4),and Yueritan-Wangjiagou Fault(F5),were significantly affected by regional main faults of ShangDan(F1) and primarily controlled by NWW and sinistral strike-slip faults. Not only did the recorded faults are extracted,but also a number of new faults were interpreted and tested,which would greatly enrich the geological structure information of this area. The faults were manly close to NWW trending,EW in some individual area,and basically presented a nature of sinistral strike-slip.