A Hidden Markov Tree Model for Flood Extent Mapping in Heavily Vegetated Areas based on High Resolution Aerial Imagery and DEM: A Case Study on Hurricane Matthew Floods
A Hidden Markov Tree Model for Flood Extent Mapping in Heavily Vegetated Areas based on High Resolution Aerial Imagery and DEM: A Case Study on Hurricane Matthew Floods
复制标题
基于高分辨率航空图像和 DEM 的植被茂密地区洪水范围测绘的隐马尔可夫树模型：飓风马修洪水案例研究
DOI：
10.1080/01431161.2020.1823514
复制
发表时间：
2021
影响因子：
3.4
通讯作者：
Sainju, Arpan Man
中科院分区：
文献类型：
--
作者：
Jiang, Zhe;Sainju, Arpan Man
洪度制图在灾害管理和国家水资源预报中起着至关重要的作用。近年来，随着大量小型卫星和无人机的部署，高分辨率光学图像变得越来越容易获得。然而，分析这些图像数据以提取洪水范围面临着独特的挑战，这是由于丰富的噪声和阴影，障碍物（如树冠，云），以及由于空间异质性导致的像素类（洪水，干燥）之间的光谱混淆。现有的机器学习技术通常侧重于光栅图像的光谱和空间特征，而没有在分类模型中充分考虑地理地形。相比之下，我们最近提出了一种新的机器学习模型，称为地理隐马尔可夫树，它以整体的方式整合了像素的光谱特征和来自数字高程模型（DEM）数据（即水流方向）的地形约束。本文通过对美国国家海洋和大气管理局（NOAA）国家大地测量局（National Geodetic Survey）的高分辨率航空图像以及DEM进行案例研究，对该模型进行了评估。在2016年飓风马修洪水期间，在北卡罗来纳州格里姆斯兰和金斯顿城市附近植被茂密的洪泛区选择了三个场景。结果表明，所提出的隐马尔可夫树模型优于几种最先进的机器学习算法（例如，随机森林，梯度增强模型），在我们的数据集上，f分数（用户精度和生产者精度的调和平均值）从70%到80%提高到95%以上。
Flood extent mapping plays a crucial role in disaster management and national water forecasting. In recent years, high-resolution optical imagery becomes increasingly available with the deployment of numerous small satellites and drones. However, analyzing such imagery data to extract flood extent poses unique challenges due to the rich noise and shadows, obstacles (e.g., tree canopies, clouds), and spectral confusion between pixel classes (flood, dry) due to spatial heterogeneity. Existing machine learning techniques often focus on spectral and spatial features from raster images without fully incorporating the geographic terrain within classification models. In contrast, we recently proposed a novel machine learning model called geographical hidden Markov tree that integrates spectral features of pixels and topographic constraints from Digital Elevation Model (DEM) data (i.e., water flow directions) in a holistic manner. This paper evaluates the model through case studies on high-resolution aerial imagery from the National Oceanic and Atmospheric Administration (NOAA) National Geodetic Survey together with DEM. Three scenes are selected in heavily vegetated floodplains near the cities of Grimesland and Kinston in North Carolina during Hurricane Matthew floods in 2016. Results show that the proposed hidden Markov tree model outperforms several state of the art machine learning algorithms (e.g., random forests, gradient boosted model) by an improvement of F-score (the harmonic mean of the user's accuracy and producer's accuracy) from around 70% to 80% to over 95% on our datasets.
登录
查看更多内容
DOI：
10.1145/3219819.3220053
发表时间：
2018-05
期刊：
Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining
影响因子：
--
作者：
Miao Xie;Zhe Jiang;Arpan Man Sainju
通讯作者：
Miao Xie;Zhe Jiang;Arpan Man Sainju
DOI：
--
发表时间：
2002-09
期刊：
--
影响因子：
--
作者：
Xiaojin Zhu;Zoubin Ghahramani
通讯作者：
Xiaojin Zhu;Zoubin Ghahramani
基于焦点测试的空间决策树学习：结果总结
DOI：
10.1109/icdm.2013.96
发表时间：
2013
期刊：
2013 IEEE 13th International Conference on Data Mining
影响因子：
--
作者：
Zhe Jiang;S. Shekhar;Xun Zhou;Joseph K. Knight;J. Corcoran
通讯作者：
J. Corcoran
DOI：
10.3390/rs11020191
发表时间：
2019-01
期刊：
Remote. Sens.
影响因子：
--
作者：
M. S. Rahman;L. Di;E. Yu;Li Lin;Chen Zhang;Junmei Tang
通讯作者：
M. S. Rahman;L. Di;E. Yu;Li Lin;Chen Zhang;Junmei Tang
影响因子：
3.9
作者：
RONEN, O;ROHLICEK, JR;OSTENDORF, M
通讯作者：
OSTENDORF, M