The Ground-objects Classification Based on Post-earthquake Airborne LiDAR Data
-
摘要: 利用机载激光雷达扫描（Light Detection and Ranging，LiDAR）技术所得点云进行震后倒塌建筑物提取时，树木与倒塌建筑物的点云特征十分相似，较难区分。为了快速准确获取震后房屋建筑物的受损情况，本文提出使用回波次数比特征指标，结合前人所提出的点云回波强度、归一化强度、最邻近点高差、法向量夹角、X向坡角和Y向坡角等特征的均值和标准差，利用K-最近邻分类法实现单体地物区分的方法。对2010年海地7.0地震震后机载LiDAR数据进行了地面点去除，分别选取了未倒塌建筑物、倒塌建筑物和树木各50个训练样本和各20个测试样本，计算了各因子的分布及其均值和标准差，在分析的基础上最终选取了可分性较强的8个分类特征，利用K-最近邻分类法对测试样本进行了分类，结果显示分类正确率可达85%以上。研究表明选取多个有效的LiDAR点云分类特征可以较好地区分震后未倒塌建筑物、倒塌建筑物和树木，提高震后建筑物震害程度判定的准确性，为应急救援及时提供较为准确的灾情信息支持。Abstract: It is difficult to distinguish the characteristics of the point cloud from the collapsed buildings, since the characteristics of the cloud and the collapsed building are similar. In this paper, we present the ratio of echo ratio and combined with the previous point cloud echo intensity, normalization intensity, altitude difference of the nearest neighbor, normal vector, X-slope and Y-slope of points, and compute the mean and standard deviation of every feature to obtain damage buildings after the earthquake. To discriminate single object, we use the classification algorithm based on K-Nearest Neighbor. We select 150 point clouds samples of 50 typical undamaged building, 50 collapsed building and 50 tree as samples from airborne LiDAR point cloud data which got after the 2010 Haiti earthquake with MW 7.0 by the way of human-computer interaction and compute the mean and standard deviation of every feature. We apply K-Nearest Neighbor to classify test samples by 8 optimal factors chosen from means and standard deviations. The classification accuracy is 85%, which indicates that the optimal factors are effective and the proposed method in this study is capable of distinguishing between building (undamaged and damaged building) points and tree points, which can extract damaged buildings and support for emergency rescue after earthquake.
-
图 5 CSF滤波法从研究区1的原始点云（a）中分离的地面点（b）结果图
Figure 5. Result of ground points (b) after separating from original point cloud (a) in the study area 1 by CSF filtering
图 6 不同地物各特征因子分布
Figure 6. Distribution of characteristic factors for different objects
图 7 各类别样本均值和标准差分布（一）
Figure 7. Distribution of mean and standard deviation of each sample
-
陈鑫连, 1995.地震灾害的航空遥感信息快速评估与救灾决策.北京:科学出版社. 程亮, 龚健雅, 2008.LiDAR辅助下利用超高分辨率影像提取建筑物轮廓方法.测绘学报, 37(3):391-393, 399. http://www.cnki.com.cn/Article/CJFDTOTAL-CHXB200803021.htm 邓非, 徐国杰, 冯晨等, 2010.LiDAR数据与航空影像结合的建筑物重建.测绘信息与工程, 35(1):35-37. http://www.cnki.com.cn/Article/CJFDTOTAL-CHXG201001021.htm 窦爱霞, 马宗晋, 黄文丽等, 2013.基于机载LiDAR和多光谱图像的建筑物震害自动识别方法.遥感信息, 28(4):103-109. http://www.cnki.com.cn/Article/CJFDTOTAL-YGXX201304021.htm 黄树松, 窦爱霞, 王晓青等, 2016.基于震后机载激光雷达点云的建筑物破坏特征分析.地震学报, 38(3):467-476. doi: 10.11939/jass.2016.03.014 姜立新, 帅向华, 聂高众等, 2012.地震应急指挥协同技术平台设计研究.震灾防御技术, 7(3):294-302. doi: 10.11899/zzfy20120308 马洪超, 姚春静, 张生德, 2008.机载激光雷达在汶川地震应急响应中的若干关键问题探讨.遥感学报, 12(6):925-932. http://www.cnki.com.cn/Article/CJFDTOTAL-YGXB200806014.htm 宿渊源, 张景发, 2012.三维激光扫描技术在震后地面三维重建中的应用.地壳构造与地壳应力文集, 0(0):30-37. http://www.cnki.com.cn/Article/CJFDTOTAL-SEIS201200007.htm 滕敏, 卫文学, 滕宁, 2015.K-最近邻分类算法应用研究.软件导刊, 14(3):44-46. doi: 10.11907/rjdk.143927 王娜, 侯爽, 2009.K-最近邻分类技术的新发展与技术改进.河北省科学院学报, 26(4):11-13. http://www.cnki.com.cn/Article/CJFDTOTAL-HBKX200904005.htm 王晓青, 王龙, 王岩等, 2008.汶川8.0级大地震应急遥感震害评估研究.震灾防御技术, 3(3):251-258. doi: 10.11899/zzfy20080308 杨乐, 曾薇, 谭颖, 2012.地震应急卫星通信系统的设计与应用.震灾防御技术, 7(1):100-109. doi: 10.11899/zzfy20120112 张小红, 2007.机载激光雷达测量技术理论与方法.武汉:武汉大学出版社. 张志友, 2008.基于LiDAR数据和航空影像的地形与建筑物提取及三维可视化.北京:北京交通大学. Dorninger P., Pfeifer N., 2008. A comprehensive automated 3D approach for building extraction, reconstruction, and regularization from airborne laser scanning point clouds. Sensors, 8(11):7323-7343. doi: 10.3390/s8117323 Dou A. X., Ma Z. J., Huang S. S., et al., 2016. Building damage extraction from Post-earthquake airborne LiDAR data. Acta Geologica Sinica, 90(4):1481-1489. doi: 10.1111/acgs.2016.90.issue-4 Fritz H. M., Phillips D. A., Okayasu A., et al., 2012. The 2011 Japan tsunami current velocity measurements from survivor videos at Kesennuma Bay using LiDAR. Geophysical Research Letters, 39(7):L00G23. http://adsabs.harvard.edu/abs/2011AGUFMNH13G..05F Hoppe H., DeRose T., Duchamp T., et al., 1992. Surface reconstruction from unorganized points. In:Proceedings of the 19th Annual Conference on Computer Graphics and Interactive Techniques. New York, NY, USA:ACM, 71-78. Han J. W. , Kamber M. , 2007. 数据挖掘: 概念与技术. 范明, 孟小峰, 译. 北京: 机械工业出版社. Labiak R. C., van Aardt J. A. N., Bespalov D., et al., 2011. Automated method for detection and quantification of building damage and debris using post-disaster lidar data. In:Proceedings of SPIE 8037, Laser Radar Technology and Applications XVI. Orlando, Florida, United States:SPIE, 80370F. Margesson R., Taft-Morales M., 2010. Haiti earthquake:crisis and response. In:DTIC Document. Congressional Research Service. Trinder J. C., Salah M., 2012. Aerial images and LIDAR data fusion for disaster change detection. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume 1-4, 2012 ⅩⅫ ISPRS Congress, 25 August-01 September 2012, Melboume, Australia:227-232. Zhang W. M., Qi J. B., Wan P., et al., 2016. An easy-to-use airborne LiDAR data filtering method based on cloth simulation. Remote Sensing, 8(6):501. doi: 10.3390/rs8060501 -
下载:
点击查看大图
图(11)
计量
- 文章访问数: 112
- HTML全文浏览量: 42
- PDF下载量: 0
- 被引次数: 0
