Research On Rapid Evaluation Method Of Road Economic Losses Caused By Landslides Based On Remote Sensing Images

In China,geological disasters occur with high density and high frequency.The number of landslides in 2010-2019 accounts for 71.33% of the total number of geological disasters.Landslides cause serious damage to buildings,lifeline system engineering and other disaster affected bodies.Roads,as infrastructure and lifeline system engineering,especially national and provincial trunk roads,have a huge impact on the national,social and economic development and people’s life.Satellite remote sensing technology can obtain information all day and all day,and make full use of remote sensing image to extract disaster information in disaster prevention and mitigation,which lays the foundation for rapid assessment of landslide emergency.The rapid assessment of the economic losses caused by landslides is helpful for the Ministry of Finance and the emergency department to make decisions on disaster relief funds,and provides reference and suggestions for the scientific layout of disaster relief work and the dispatch of disaster relief resources.By comparing the process of economic loss assessment in disaster field investigation and combining with the research status at home and abroad,this paper puts forward a rapid assessment method of road economic loss caused by landslide based on remote sensing image.Firstly,according to the spectral,geometric and contextual characteristics of landslides and roads in remote sensing images,the interpretation marks of different damage levels of roads are established,including "basically intact","partially damaged" and "damaged".Secondly,the method of extracting landslide and road information based on remote sensing image is studied.Combined with the object-oriented classification method,the image segmentation,feature extraction and damaged road identification of road buffer area and disaster area are carried out.The damaged location,damaged length and landslide vector range of different damage levels are interpreted.Then,the external geometry of landslide is abstracted as ellipsoid,wedge and cone,and the volume of each landslide point is obtained by using the corresponding volume calculation formula.The direct economic loss includes the cost of cleaning up the landslide and the economic loss of destroying the road,and the corresponding calculation formula is given.The indirect economic loss caused by the landslide is estimated by using the empirical coefficient method,and the total economic loss caused by the landslide is accumulated.In this paper,the landslide in Ji Re section of national highway G216 in Ji Long County after Nepal earthquake on April 25,2015 was verified.The total classification accuracy was 0.87,kappa coefficient was 0.77,and the classification quality met the high consistency.Based on the object-oriented classification,the road damage information and landslide vector range are extracted,and the error rate between the geometric model estimation and the measured landslide volume is between-24.09% and 25.70%.Compared with the landslide volume estimated by the empirical model in the statistical regression formula method,the error rate of the geometric model estimation is in a reasonable range.Finally,the direct economic loss of Ji Re section of national highway G216 in Ji Long county is 104.8446 million yuan,and the total economic loss is 1153.2911 million yuan.