The Change Detection Of Buildings In Seismic Damage Based On Multi-temporal And High Resolution Remote Sensing Image

With the continuous development of remote sensing technology, the era of remotesensing data acquisition comes to a new stage in which the way of data acquisition emergesinto a multi-platform, multi-sensor, multi-resolution (multiple spatial resolution,multi-spectral resolution, multi-time resolution). Because of the constantly increasing datatype and the improving of the data quality, the study on change detection has become morefrequently and used in many fields based on high resolution remote sensing images. Atpresent, various devastating natural disasters occurred frequently, which caused hugecasualties and property losses. Disaster analysis which using multi-temporal high resolutionremote sensing images has become a hot research at present. The changes of constructionafter the earthquake were studied in this paper, and which will take an important role for theemergency rescue of post-earthquake and post-disaster reconstruction work. The specificcontents in this paper can be described as follows:1) Firstly, the characteristics of high resolution remote sensing images were analysed andsummarized in this paper, and the related characteristics of the image were found out for thisstudy. Secondly, the principle、advantages and disadvantages of common remote sensingchange detection methods were analysed and described in detail, including Image Arithmetic,Vegetation Index Differencing, Principle Component Analysis and Post ClassificationComparison.2) Studying the object-oriented classification techniques, which combined the principleof multi-scale image segmentation and fuzzy classification of image.3) A sequential change detection method base on several characteristic was proposed inthis paper, the main steps of the method are: first doing the pixel-level change detection toachieve candidate region including buildings change; Then to detect the change of texturefeatures in the candidate region; At last to detect the change of color feature based ondetection about texture feature, so with this method, it not only reduced the range of changedetection, but also reduced the interference of other objects step by step. In order to improvethe efficiency of the method, the preprocessing of the data must be carried out before thechange detection. Meanwhile the principle and method of radiation correction and geometriccorrection were introduced in this paper.4) A method based on overlay analysis and statistics in the same class about changedetection. Using the object-oriented image analysis technique and overlay analysis in spaceanalysis based on vector data. First extracting buildings by image classification, then findingthe area of damaged buildings by overlay analysis and assessing the state of buildings in earthquake by correlation coefficient, finally getting information about the level of damage ofbuildings.At last experiments were done, with the data of some of the affected area of a Ms7.1earthquake in Yushu county of the Qinghai province in2010. The experimental results showthat, the two methods proposed in this paper is effective to improve the correct rate ofbuilding change detection, and reduce the false detection rate.