Research On Shadow-based Building Extraction From High Resolution Remote Sensing Images

In recent years, with the rapid development of aerospace remote sensing technology. Satellite remote sensing images resolution has been significantly improved. Extracting terrain information from remote sensing images also showed a short cycle, fast, informative and so on. Building in remote sensing images is the most fundamental and most important information in geographic information, and also the basis of data in data collection of Geographic information system. Information from the building on the GIS data acquisition, image understanding, mapping and spatial database updates, etc. have a far-reaching and important. At the same time, the existing building extraction from high resolution remote sensing method is not automated, background identified as target objects or target objects identified as the background and so on. Therefore, how to accurately, low false positive extract building in high resolution remote sensing images has become the important issues of remote sensing interpretation technology.Mathematical morphology is an important aspect of the Image processing theory. It has the characteristics of simple operation, simplify image data, maintaining the basic shapes of the image which has been widely used in image enhancement, noise filtering, the enhance and skeleton of shape and structure, shape reconstruction, image compression and so on. This makes the morphological structures in the extraction of high resolution image has a unique purpose and advantages.Probability spatial relations exist between buildings and their shadows,which means that buildings can be extracted by shadow detection. In order to improve the accuracy of building extraction,it should detect vegetation area and find the trees shaded area caused by tall trees through the knowledge of color space after detection of the shaded area shadow region by the characteristics of the shadow. By overlapping the two regions and then erase the shaded area of trees to find the building shadows. Use the spatial relation information between building and its shadows,the author identify the seed point in the building area and extracting buildings through the region growing method.The high-resolution remote sensing images in this paper are from Google Earth and using Visual Studio 2008.net software to achieve its effect. Manual extraction results compared with experimental study results of design of real and the same scene shows that extraction of Buildings in the final results has high accuracy. By this thesis, the shadow-based building extraction method of high resolution remote sensing image is feasible and effective.