高分辨率遥感影像居民地半自动提取方法研究

The method for semi-automatic extraction of residential area from high resolution remote sensing images

居民地是地震灾害的重要受灾体,是进行地震应急救援准备的基础。采用高分一号2 m分辨率卫星遥感数据提取研究区真实居民地空间分布信息,为地震应急救援准备提供数据支持。结果表明:运用灰度共生矩阵、二值化、数学形态学等方法对影像进行迭代运算,能较好地提取高分一号2 m分辨率影像居民地信息。这种算法正确率较高,鲁棒性较好,然而对于稀疏植被及非居民聚居的人工建筑区提取结果虚警率因子较高,存在误提现象,需将提取结果与影像进行对比分析,通过人工干预完成研究区居民地半自动提取,以保证提取精度。

Residential area,as an important hazard-bearing body of earthquake disasters,usually constitutes the basis of earthquake emergency preparation. In this paper,2 m resolution GF-1 satellite remote sensing data were used to extract the spatial distribution information of residential area,which could provide data support for the earthquake emergency preparation. The results reveal that more accurate residential area information of the high resolution GF-1 2 m image can be extracted based on gray level co-occurrence matrix,binarization and mathematical morphology. This proposed algorithm has high accuracy and good robustness. However,higher false alarm rate factor was shown in the extraction of sparse vegetation and non-residential buildings. Thus,in order to ensure data accuracy,the authors compared and analyzed the extraction results and the images,and extracted the ultimate data results semi-automatically by the artificial intervention.