结合PolSAR影像纹理特征分析提取倒塌建筑物

发布时间：2018-08-05 20:26

【摘要】：利用震后1景极化SAR影像提取倒塌建筑物是一种快速有效的灾害调查手段。倒塌建筑和倾斜建筑物在PolSAR影像中的散射特征相似,易造成建筑物倒塌率的过度评估。由于倒塌建筑和倾斜建筑的纹理特征有较大差异,将利用这种纹理差异来解决倒塌建筑和倾斜建筑的混分问题。通过实验发现均值、同质性、熵及相关性4种基于灰度共生矩阵(Gray-Level Co-occurrence Matrix,GLCM)的纹理特征能够有效区分倾斜建筑和倒塌建筑,故利用这4种纹理特征提取倒塌建筑中混杂的倾斜建筑,从而降低倒塌建筑的虚警率。以玉树地震为例,提取城区的建筑物震害信息,实验证明该方法能够大幅提高建筑物震害评估精度。
[Abstract]:It is a rapid and effective method to extract collapsed buildings by using 1-scene polarized SAR image. The scattering characteristics of collapsed buildings and inclined buildings in PolSAR images are similar, which can easily cause the over-evaluation of collapse rate of buildings. Because the texture characteristics of collapsed building and inclined building are quite different, this kind of texture difference will be used to solve the problem of mixing collapsed building and inclined building. It is found by experiments that four texture features based on Gray-Level Co-occurrence matrix (Gray-Level Co-occurrence Matrixer GLCM) can effectively distinguish the inclined building from the collapsed building, so the four texture features can be used to extract the mixed inclined building in the collapsed building. In order to reduce the false alarm rate of collapsed buildings. Taking the Yushu earthquake as an example, the earthquake damage information of buildings in urban area is extracted, and the experimental results show that the method can greatly improve the accuracy of building damage assessment.
【作者单位】：中国科学院西北生态环境资源研究院;中国科学院大学;中国科学院西北生态环境资源研究院甘肃省遥感重点实验室;中国科学院西北生态环境资源研究院黑河遥感试验研究站;甘肃省地震局;武汉大学资源与环境科学学院;
【基金】：中国科学院“百人计划”项目“寒旱区地表水文关键要素的多源遥感数据同化研究”(29Y127D01) 甘肃省地震局地震科技发展基金“基于极化雷达的建筑物震害识别研究”(2015M02)资助
【分类号】：P237;P315

【相似文献】