Research On Automatic Extraction Method Of Alluvial Fan Gravel Coverage Area Based On GF-1

Nearly two-thirds of the land in China is hilly and mountainous.The rushing and flooding fans are mostly distributed in front of the mountain.It is closely related to people's life and production.At the same time,the alluvial fan and the debris flow are closely related to each other,and it is easy to landslide and Geological disasters such as mudslides,and the distribution of scouring and alluvial fans usually have loose soils,and their distribution range has a great influence on engineering geological site selection.In oil and gas,the favorable alluvial fan reservoirs can be inferred through the exploration of surface alluvial fans.Therefore,it is important to quickly determine the distribution range of rushing and flooding fan gravel on remote sensing,which is of great significance for geological disaster prevention,oil and gas exploration and engineering geological site selection.The paper studies the application of high score satellite data in the automatic extraction of alluvial fan gravel coverage.Based on the characteristics of alluvial fan and the characteristics of the study area,an alluvial fan gravel distribution model "FGDM"(Alluvial Fan Gravel Distribution area extraction Model)is proposed.The basic principle is to select the three indexes of the Alluvial Fan index AFI,the NDSI and the NDVI as the main factors.The AFI increases the spectral difference between the alluvial fan and other features through the ratio of the red and near red bands.NDSI highlights sand information;the NDVI index excludes vegetation areas.The three index calculation results are combined into a new multi-band image,and the principal component change is selected to select the first principal component.Then use the histogram threshold to determine the gravel coverage area of the alluvial fan cluster,and use mathematical morphology filtering to remove fine spots and voids;then,extract the inter-fan concave land between the alluvial fans through the study area DEM data,and finally determine the single alluvial A range of boulder-covered areas.Taking the key alluvial fan in Heshuo area of Bayinguoleng Mongolian Autonomous Prefecture in Xinjiang as an example,the paper adopts FGDM to extract it based on automatic registration and image fusion.In addition,as a comparison,it also combines Visual interpretation and field work analyze the extraction accuracy of the alluvial fan.Finally,the extraction results are compared with field verification and visual interpretation.The alluvial fan extracted by FGDM is consistent with the results of visual interpretation and field verification.Through the out-of-field verification and visual interpretation comparison analysis,it can be concluded that the automatic extraction method designed in this paper is feasible and effective,and realizes the division of a single alluvial fan boundary in the alluvial fan group.The technology has two major innovations.One is to propose a comprehensive index for the extraction of alluvial fan gravel coverage areas,and the other is to improve the extraction accuracy of alluvial fan gravel coverage areas.