Research On Change Detection Of Seismic Disaster Information Extraction For Remote Sensing Images

The earthquake is one of the most serious natural disasters,it has the characteristics of sudden,short duration and destructive,which constitutes a great threat to the safety of people's lives and property.The traditional earthquake emergency and seismic disaster information need to be collected artificially,which cannot achieve the earthquake emergency rescue demand,leading to rapidly increase of mortality.With the emergence of high resolution remotely sensed images and development of remote sensing information extraction technology in recent years,remote sensing gradually became an effective means for acquiring seismic disaster information rapidly,emergency response and seismic disaster assessment.The main works of this thesis are summarized as follows:1.Through the observation of buildings in high resolution remote sensing images,Spectra,geometric and texture features of buildings are thoroughly analyzed on the earthquake pre-and post-earthquake image.Through theoretical analysis and experimental results,the image parameters of the separation of buildings and other objects are gained.2.The multi-scale segmentation algorithm based on the minimum principle of heterogeneity is analyzed.The optimal segmentation scale is presented using the global spatial autocorrelation Moran index and the standard deviation of the image object.Furthermore,the effectiveness of algorithm is verified by experimental data.3.The object-oriented classification is analyzed and the SVM-KNN classifier is introduced.The principle and the concept of SVM and KNN classifier are described.By pointing out the shortcomings of SVM,the combination of KNN and SVM is adopted to compensate the deficiency of the classification accuracy of a single classifier.Through building extraction and classification experiment,the classification of buildings pre-and post-earthquake image changes detection,to extract the change area.4.Through a series of experiments on pre and post-earthquake remote sensing image analysis,the system based on MATLAB software platform is designed in this paper,mainly including multi-scale segmentation,classification and change detection modules,and experimented and verified effectiveness of the platform and application of the algorithm.Through the destruction of buildings,the casualties of the disaster area can be judged according to the damage of buildings,and the earthquake intensity can be mapped quickly.