Font Size: a A A

Application Of Refined Terrain Modeling Based On UAV Image In Landslide Recognition

Posted on:2020-09-27Degree:MasterType:Thesis
Country:ChinaCandidate:J L LiuFull Text:PDF
GTID:2370330572980195Subject:Geological engineering
Abstract/Summary:PDF Full Text Request
With the expansion of the scope of human activities and the continuous advancement of engineering construction,the frequency of landslide disasters is on the rise,people's lives,property and economic construction has also been a great threat.Therefore,the use of fast,efficient and accurate landslide extraction means to identify potential landslide,timely determination of the location of the landslide,the scope of influence,targeted measures,can reduce the impact of landslide disaster on human beings,disaster prevention and mitigation and the next step of planning and construction is of great significance.Although domestic and foreign scholars have carried out more research on landslide remote sensing,the application potential of remote sensing technology has not been fully excavated,which is limited by the development of cross disciplines such as funds,data acquisition,processing and information extraction technology,remote sensing geological interpretation and so on.The key problems in the interpretation of landslide disasters in the past are insufficient,mostly focusing on the spectral reflectivity,the shape of the slope body and other characteristic information and the cumulative deformation of the landslide,the mathematical method is not related to the landslide characteristics,can not realize the landslide geoscience theory knowledge and expert experience to transfer to computer recognition,and the ground resolution of remote sensing images is not high,landslide remote sensing in the interpretation of the one-sided pursuit of spectral reflection characteristics,DEM,The underutilization of data such as contour lines,especially the occurrence of landslide disasters is sudden,the environment of disasters is more complex,in the harsh geological environment,landslide disaster information is difficult to obtain in a timely and effective manner,the past technical means in landslide extraction and treatment has certain limitations.Based on the research topic of fine terrain modeling of UAV images and its application in landslide extraction,this paper takes the Jiang jia gou River basin in Dong chuan,Yunnan Province as an example,and obtains the following research results in fine terrain modeling and landslide extraction in Highland mountainous areas:(1)According to the principle of aerial photogrammetry of UAV,the source of error is analyzed in detail,and according to the characteristics of aerial photogrammetry of UAV,a set of UAV aerial photogrammetry error control scheme is developed,which provides the necessary technical support for obtaining high precision image data.(2)Aiming at the problems of complex terrain and fine terrain modeling errors in Highland mountainous areas,this paper introduces UAV aerial photogrammetry technology,and through the research of UAV aerial photogrammetry technology,establishes a set of methods and technical system suitable for fine terrain modeling in highland mountainous areas,The image data of 0.14 m spatial resolution in the study area is obtained,and the DOM with planar point error of 0.547 m is constructed,and the DEM with the error of 0.684 m in elevation provides the necessary data basis for further extraction and quantitative analysis of landslide.(3)According to the previous study of landslide extraction,the application of geoscience knowledge and expert experience in landslide extraction is not sufficient,the primary and secondary of landslide extraction index is not clear,this study introduced mathematical statistical analysis distance discrimination method,combined with the existing geological data,the establishment of two landslide extraction indicators,According to the 6 main landslide extraction indexes extracted from the fine terrain model of UAV image,the discriminant model of landslide is constructed,and the miscalculation rate of learning sample and test sample is 6.7% and 0 respectively,which has high recognition accuracy,which is simple,fast and accurate,and has high application value in the process of identifying landslide disaster.
Keywords/Search Tags:UAV images, error control, terrain modeling, mathematical statistics judgment method, landslide extraction
PDF Full Text Request
Related items