高分辨率遥感影像居民地半自动提取方法研究

发布时间：2018-02-23 23:47

本文关键词： 高分辨率遥感高分一号居民地　出处：《国土资源遥感》2017年03期 　论文类型：期刊论文

【摘要】：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。
[Abstract]:Residents' land is an important disaster body of earthquake disaster, and is the basis of earthquake emergency rescue preparation. The spatial distribution information of real residents in the study area is extracted by satellite remote sensing data with high resolution of 1 ~ 2 m. The results show that the gray level co-occurrence matrix, binarization, mathematical morphology and other methods are used to iterate the image. The algorithm has higher accuracy and robustness, but the false alarm rate factor is higher for sparse vegetation and non-resident artificial building areas. There is a phenomenon of false extraction, so it is necessary to compare and analyze the extraction results with the image, and to complete the semi-automatic extraction of residents in the study area by artificial intervention to ensure the extraction accuracy.
【作者单位】：新疆维吾尔自治区地震局;中国地震局地质研究所;
【基金】：地震应急青年重点任务“面向地震应急准备的居民地数据提取研究”(编号:CEA_EDEM-201511) 新疆地震科学基金项目“三维影像技术在承灾体基础数据调查中的应用——以吐鲁番市主城区为例”(编号:201510)共同资助
【分类号】：TP751

【相似文献】