Research On The Evaluation Method Of Housing Seismic Rating Based On Remote Sensing Technology

Earthquake is one of the most serious natural disasters facing mankind at present,and as the main body of earthquake disasters,housing construction has a direct impact on the direct economic losses and casualties caused by earthquakes.Therefore,it is of great significance to fully grasp the spatial distribution of existing housing buildings and their seismic resistance,which is of great significance for finding out the risk base of earthquake disasters,predicting earthquake damage losses,and formulating emergency rescue strategies.The assessment of the seismic resistance of traditional houses is mainly from the perspective of civil engineering structure,and the field investigation of the structural type,floor height,wall material,roof type,foundation foundation and seismic structural measures of the house is carried out,and the seismic resistance of the house is comprehensively evaluated.The traditional method of evaluation is accurate,but due to the high professionalism of the assessment,the large workload and the low efficiency,it is difficult to carry out on a large scale.The use of remote sensing images to judge the seismic resistance of the house can improve the efficiency of the seismic ability of the house.Based on the deep learning remote sensing image processing algorithm,this paper proposes a multi-scale aggregation full-convolutional neural network automatic extraction method for housing objects,and realizes the preliminary judgment of semiautomatic/semi-intelligent housing seismic resistance level based on remote sensing images and empirical estimation methods.The main research contents and achievements of this paper are as follows:（1）On the basis of comprehensive analysis of the current remote sensing image classification method,a multi-scale aggregate fullconvolutional neural network for automatic extraction of houses is proposed based on the deep learning framework,and the automatic extraction of housing areas and singlefamily house contours on high-resolution remote sensing images is realized.（2）Taking Hubei Province as an example,on the basis of the traditional housing seismic evaluation standards,the evaluation standards for the seismic rating of houses suitable for remote sensing images are established,and the sample library of typical housing remote sensing images with different seismic ability levels is established to provide a basis for human-computer interactive interpretation.（3）The automatic extraction of highresolution remote sensing images of houses in Hubei Province was carried out,and a total of more than 10.6 million single-family houses were extracted.Using humancomputer interaction interpretation method to complete the classification of the seismic ability of houses in Hubei Province,and the preliminary judgment of the seismic ability of houses,the experimental results show that the estimated seismic resistance of Hubei Province is about 35.3%,the houses suspected of having substandard seismic ability are about 47.7%,and the houses with suspected earthquake resistance are seriously insufficient are about 17.0%.The results of field investigation and error analysis prove the effectiveness and accuracy of this method,and at the same time illustrate that the proposed method is suitable for large-scale housing seismic resistance survey,which can effectively improve the investigation efficiency.