高分辨率遥感影像的形态学房屋指数模型

he space observation technology which combined with high-resolution images develops rapidly,and plays a significant role in urban development evaluation, land use and other aspects.As the fundamental elements of the city, identification and extraction of buildings are important to urban planning, urban monitoring and other aspects.In this paper, mathematical morphology is studied. By considering the characteristics of the buildings in the high-resolution image such as brightness, shape, directionality, etc, we proposed a novel morphology building index (MBI) for automatic building extraction based on mathematical morphology, used morphology operators as granulometry, differential, top-hat, reconstruction. The best characteristic of MBI is that MBI algorithm can quickly, precisely and automatically extract buildings. Therefore, MBI has a great value in urban monitoring, big data information filter, etc. Based on the MBI model, large amount of images have been processed, it stated that MBI can automatic and accurate for building extraction. Such as: IKONOS Harbor of Australia, Worldview-2 Wuhan, Hangzhou, QuickBird Wuhan, etc, all can give more than 85% overall accuracy,Worldview-2 Hangzhou dataset even can obtain 90% overall accuracy. In this paper, improved MBI algorithms such as: max algorithm, directional algorithm have been proposed, complete the MBI model, and also have great accuracy.