Building Areas Detection From Optical Remote Sensing Images Based on Hierarchical Structure Description

Automatically detection of suburban building areas(BA) is one of the research hotspots in the field of remote sensing image(RSI), which is widely used in land monitoring and anti-terrorism. However, due to the large differences in the appearance of suburban buildings, the mainstream methods are difficult to obtain better detection results. To address these problems, this paper presents an hierarchical structure description suburban building areas detection framework, which can be divided into two main stages: (1)in the stage of extraction of key structure(s)(KS), we propose a candidate screen method based on roof parallel structure(RPS), to achieving rapid extraction KS. (2)in the stage of candidate region identification, we present a Deformation Adaptability Model, this model can make full used of the relationship between key structures of buildings, and it has better adaptability to distributed target. We test the proposed method on both ordinary and complex structure data sets, and the experimental results demonstrate that our framework can achieve high accuracy and efficiency.