Study On Method Of Shadow Detecting And Extraction From Quickbird Imagery

High-resolution satellite images such as IKONOS,Quickbird,OrbView and Worldview have played a more important role in our life.We can extract more detailed information on urban environment,precision agriculture,traffic facility and military target from high resolution images.Though high resolution remote sensing images provide us richer information of surface targets,how to identify and extract features from these images automatically and quickly has been a big challenge.Shadow is a usual problem in remote sensing images,which blocks part of the surface information.It is particularly obvious in high spatial resolution satellite imagery.Especially in dense urban area,the shadows of tall buildings block a lot of surface information,those of the lower buildings also screen a lot of information of roads around.The treatment of the shadow problem in high-resolution satellite images is particularly important.In this thesis,we try to extract the edge of the shadow in QuickBird imagery.Based on actuality of the theoretical basis and detection methods of the shadow,DB wavelet analysis method,combined with the texture characteristics of QuickBird satellite imagery is used to detect shadow.The full text is divided into five parts.The first chapter discusses the research of this thesis,and then analyzes the state-of-art technology and trends of shadow identification from high-resolution remote sensing images.And explains the thesis structure and organization finally.ChapterⅡfocuses on the characteristics of the shadow in remote sensing imagery and fulfills a comprehensive analysis of the shadow edge character and the spectral features. ChapterⅢreveals texture features of the Quickbird image,and summarizes the existing methods of texture analysis and their deficiencies.ChapterⅣdiscusses the existing methods of wavelet analysis,and then uses the DB wavelet basing on Quickbird image texture analysis with the different windows,then verifies the accuracy of edge extraction and evaluates the precision.After many comparing tests,the more suitable windows for extracting shadows are defined.ChapterⅤmakes main conclusions of this study and discusses the prospects of shadow detection.