Research And Application Of Extracting Building Anti-seismic Factors Based On Remote Sensing Technology

Investigating the anti-seismic factors of buildings(seismic factors)to predict the damage ratio of buildings under earthquake scenarios is an important basis for regional earthquake disaster reduction plans and seismic design standards.Due to the impact of the seismic design concept of "protecting life",the traditional single building seismic capacity verification focuses on the physical response mechanism and damage probability model.By weighting the seismic factors which affect the building damage rate such as "structure,height,fortification level",a lot of models were established to predict earthquake damage to regional buildings.In recent years,the performance-based seismic design of buildings has been proposed,the consideration of social aspects such as function,casualties,economic losses of buildings were added to the goal of seismic design.It is a new trend in the study of earthquake scenario construction and seismic loss pre-assessment,that research on not only the traditional seismic factors such as "structure,height,fortification level",but also the factors that have significant differences in spatial distribution such as "waterproof standard,population density,and economic density".The first national natural disaster risk survey put forward the demand of "wider scope and higher efficiency" for the seismic factor data survey of buildings in China.It is the main difficulties for seismic damage prediction of group buildings that the building foundation database is incomplete,the traditional field investigation method is difficult,high cost and low efficiency.How to establish a fast,efficient and low-cost seismic factor investigation method for a large scale of buildings,and to establish a seismic damage preassessment process suitable for a large spatial scale,are the key scientific issues that need to be solved urgently in natural disaster risk survey in China.In order to solve the problem that it is difficult to obtain the buildings’ parameters for seismic damage prediction,this paper studies the extraction method of regional buildings’ seismic damage factors by all kinds of remote sensing images and big data of web.Combined with the light remote sensing data,the spatial and temporal distribution characteristics of the population and economy were statistically analyzed,and the process and method of earthquake damage pre-assessment based on remote sensing were constructed.The main achievements are as follows:Firstly,based on the building seismic damage grade and seismic damage index,the main factors affecting the earthquake damage of buildings are studied.then the relationship between remote sensing indexes such as "spectral feature,texture feature,phase feature" and the seismic factors such as "profile,height,age" are analyzed.Then introduced the extraction principles of various seismic factors by remote sensing.Based on the probabilistic distribution characteristics of seismic factors which were extracted from remote sensing(remote sensing seismic factors).At last,we propose a preevaluation method of building damage ratio based on remote sensing seismic factors.Secondly,the process and method of extracting seismic factors of inter-provincial buildings are constructed based on the remote sensing data.Combined with Web big data,the K-means image classification algorithm was improved,and a rapid extraction technology for building areas combined with network data and remote sensing data was proposed.The extraction accuracy can reach more than 90%.In response to the problem of incomplete domestic building attribute databases,the " The method for obtaining seismic resistance factors of buildings in three-level sampling of cities,towns and rural areas has evaluated the characteristics of the distribution rate of building attributes of the“census data” and “1% population sample survey data” over time,with a two-sided accuracy of 99% confidence interval Within 10 years,there has been no significant change in the structure of building properties in all provinces across the country.In the end,it is believed that the distribution probability of building attributes in the "census data" can be used as the regional building seismic factor parameter input to the regional building damage ratio pre-evaluation model.Thirdly,a process to get the single building factors by remote sensing technology was established.Whith the machine learning model which named Seg Net,a building contour extraction method which based on GF2 image was established,the average extraction accuracy is 92.14%.The phase residual of Persistent Scatterer Interferometric Synthetic Aperture Radar(PS-In SAR)is introduced to extract the high-precision building height.The average value of the extraction error is-0.06 layer,and the mean square error of the error is 2.009 layer,which can meet the requirements of the building vulnerability curve evaluation model.The method of building age change detection based on time series remote sensing image and web big data is established.The recognition accuracy is46.15% 1990s-pre,63.55% in 1990 s,84.23% from 2000 to 2010,and 90.91% 2010-post.Based on ps-insar technology,we also explore the characteristics of thermal barrier and cold shrinkage of buildings.A structural identification method based on the coefficient of thermal expansion and cold shrinkage of buildings was established,and the accuracy of the structure identification can reach 70.17%.Finally,based on fuzzy evaluation method and building vulnerability curve,remote sensing technology is applied to pre-assessment of building earthquake damage in the capital circle and Sichuan province.Based on the current seismic factor data of buildings in the capital area,the Tangshan earthquake in 1976 is reproduced.The damage ratio of buildings in the area is estimated,and the building seismic index at the county scale and the simulated macro seismic intensity are produced.Based on the remote sensing data before and after the 2008 Wenchuan earthquake in Sichuan Province,the building damage,casualties and economic losses in the epicenter counties of Sichuan Province are simulated,and the accuracy of the proposed method for earthquake damage prediction is verified.