Study Of Earthquake-induced Building Damage Information Extraction By Using Change Detection Of SAR Images Coherence

When an earthquake that can bring disasters to people and buildings occurs,the country where the disasters occurred and the government directly under the jurisdiction of the affected countries must carry out rescue work.Before carrying out rescue work,the primary task is to conduct disaster assessments and accurately assess the disaster areas.If the division is good,the destruction is severe and where the losses are lighter,so that when allocating people to the disaster area to save people,it can be targeted to allocate personnel to each area.In this step of the work,remote sensing technology can be done because of the advantages of fast,large area,accurate and direct access to target area images,especially in recent years.The work includes: providing the image data taken before the earthquake occurred in the damaged area,taking the image of the disaster area after the earthquake,and using remote sensing image to evaluate the disaster loss in the damaged area.Generally speaking,the high resolution optical remote sensing data can basically meet the disaster relief needs,but when the disaster area is accompanied by cumulus clouds,fog or heavy rain and other extreme weather,the advantages of optical remote sensing are greatly limited.The SAR in microwave remote sensing is not only unaffected by such bad weather conditions,but also can acquire images normally at night as in the daytime,and when SAR is used for shooting and detecting,it is photographed from the oblique top of the object under study,this imaging work method can detect the side information of the feature more comprehensively and more meticulously.SAR image data is not only the strength of data,but also contains the phase data,this is a great advantage compared to the SAR data of the optical remote sensing data.This research is phase coherence based on SAR images of change detection,extraction of earthquake damage to buildings,the main results are as follows:(1)By analyzing the working principle of SAR sensors,the scattering effects of normal regular buildings on SAR emission electromagnetic waves and their characterization on SAR images,such as image overlay,shadow and secondary scattering,etc,are summarized.The damage characteristics of buildings on high resolution and low resolution SAR images are also analyzed and described.The purpose of this paper is to distinguish between intact and damaged buildings in earthquake damaged areas simply and directly.Therefore,only two types of buildings,"intact" and "destroyed",were extracted from the study.(2)Based on the coherence algorithm of electromagnetic wave,two sets of SAR image data in the two experimental areas are calculated and processed before and after the earthquake,respectively.Based on the coherence results,the appropriate threshold calculation method is selected to calculate the classification threshold and extract the damaged buildings.(3)Because of the existence of many kinds of decoherence factors in SAR sensors,these factors influence the results of coherence calculation to some extent.Therefore,based on the idea of vegetation growth index calculation in remote sensing of vegetation,the preearthquake and coseismic coherence normalization algorithm is introduced to detect the change of coherence and then extract the damaged buildings.(4)Based on the idea of eliminating the influence of decoherence factors on coherence,and the research of using hypothesis test idea to detect the change of optical remote sensing images to extract land use.The model of t hypothesis test is introduced to detect the change of coherence before and after earthquake.