High Resolution Remote Sensing Image Building Disaster Detection

n recent years, frequent natural disasters have caused serious damage to people’s life and property safety. The traditional ground survey methods are time-consuming, low efficiency, and difficult to meet the needs of rapid response to natural disasters. With the rapid development of remote sensing technology, people have paid a great attention to using remote sensing technology to investigate disaster information. Building detection in disaster area is significantly important for collecting the disaster information and implementing post disaster rescue.Aiming at detecting building in disaster area from high-resolution remote sensing image, we propose an improved multi-directional and multi-scale segmentation algorithm combined with morphological features, based on morphological analysis theory in high-resolution remote sensing images, to detect buildings in disaster area automated. Firstly, we integrate the properties of morphological operators (e.g., reconstruction, granularity, and directionality) into the implicit characteristics of buildings (e.g., brightness, size, and contrast) to extract bright, high-contrast buildings. Then, we combine the regional image edge information to extract potential buildings. the proposed method have a higher detection rate and a low false rate in detecting buildings in disaster area which is validated by experiments on high-resolution remote sensing images.Moreover, for the monochrome urban remote sensing images which the edge information of building is not significant, we proposed a building detection method based on building morphological profiles combined with building shadow feature. Firstly, we detect those bright, high-contrast buildings and their shadow, using multi-directional and multi-scale segmentation algorithm based on morphological profiles. Then calculate the direction and distance constraints of building shadow, to proof the building area. Proved by experiments on urban remote sensing image, the proposed method shows better detection effect.