基于高分遥感影像的地震受灾建筑物提取与倒损情况快速评估

发布时间：2018-01-02 11:18

本文关键词：基于高分遥感影像的地震受灾建筑物提取与倒损情况快速评估　出处：《自然灾害学报》2016年03期 　论文类型：期刊论文

【摘要】：以鲁甸地震为例,基于灾后无人机航拍影像,利用均值漂移分割算法以及C5.0决策树分类算法,提取了受灾建筑物区域,并对建筑物倒损率与地震烈度进行了估算,经目视解译验证以及与官方发布地震信息比对,倒损率相对误差均小于5%,地震烈度估计则与真实信息完全吻合。同时,利用采样方法计算了真实倒损面积与分类倒损面积的比值系数,提出了对于同一地震、相似地貌、相似建筑物分布的不同区域的建筑物倒损面积评估模型,经验证,其相对误差最小可低达5.31%,从而实现了倒损面积更准确的快速评估。
[Abstract]:In the case of Ludian earthquake disaster, the UAV aerial images based on segmentation algorithm and C5.0 decision tree classification algorithm using mean shift, the affected area and extraction of buildings, building on damaged rate and earthquake intensity was estimated by visual interpretation and verification with the official release of seismic information on down the loss rate of relative error are less than 5% and the seismic intensity estimation is fully consistent with the true information. At the same time, the ratio of coefficients by sampling method and classification of the damaged area of the real calculation of damaged area, is proposed for the same earthquake, similar landforms, similar buildings distributed in different regions of the building collapsed and damaged area after verification, the evaluation model. The minimum relative error is as low as 5.31%, in order to achieve a rapid assessment of damaged area more accurately.

【作者单位】：中国地质大学(北京)信息工程学院;中国科学院遥感与数字地球研究所遥感科学国家重点实验室;
【基金】：国家高分辨率对地观测系统重大专项(03-Y30B06-9001-13/15-01) 中科院重点部署项目(KZZD-EW-07-02) 国家高技术研究发展项目(2013AA12A401)~~
【分类号】：P315.9;TP751
【正文快照】： Supported by:The National Major Program on High Resolution Earth Observation System(Grant No.03-Y30B06-9001-13/15-01);the KeyResearch Program of the Chinese Academy of Sciences(Grant No.KZZD-EW-07-02);the National High Technology Researchand Development

【相似文献】