基于多纹理特征融合的震后SAR图像倒塌建筑物信息提取

合成孔径雷达（SAR）凭借其全天候观测能力以及SAR图像中丰富的纹理信息,在震后建筑物倒塌评估中发挥了重要作用。针对SAR图像中倒塌建筑物纹理特征多样但利用率较低,且特征信息冗余的问题,提出一种基于主成分分析的SAR图像多纹理特征分类方法。该方法基于灰度直方图、灰度共生矩阵、局部二值模式、Gabor滤波器提取了26种纹理特征信息,构建主成分变量进行多维特征优选与降维融合,通过随机森林分类算法提取建筑物的倒塌信息。以2016年日本熊本地震为例验证了该方法的有效性,结果显示其提取精度高达79.85%,倒塌建筑物的识别效率有所提高,分类结果优于单种纹理特征提取方法及多种纹理特征组合提取法,可用于震后建筑物震害信息的快速提取。     

极化SAR建筑物震害信息识别研究方法综述

合成孔径雷达具有全天候、全天时的对地观测优势,全极化合成孔径雷达（PolSAR）能够同时获取4种极化信息,利用PolSAR在震后进行震害评估具有及时性和准确性的优势。首先概述了PolSAR的发展状况及其在建筑物震害信息提取中的应用;其次,基于不同数据类型（多时相数据、多源数据、单时相数据）,概述了近10 a来 PolSAR数据在建筑物震害提取中的应用及其对比分析;然后,从极化分解方法、极化特征以及纹理特征3个方面对PolSAR数据的建筑物震害提取方法进行了详细阐述;最后,提出未来研究工作的设想,以期结合地理信息数据POI补充PolSAR在震害评估精度方面的不足。     

综合灰度与纹理特征的高分辨率星载SAR图像建筑区提取方法研究

高分辨率SAR图像的纹理特性对于图像的解译及地物分类等具有重要的意义。根据高分辨率星载SAR图像上建筑区的纹理有别于其他地物的特点,提出了一种综合利用灰度和纹理特征的高分辨率星载SAR图像建筑区提取方法。首先对SAR图像进行斑点噪声的抑制,然后利用灰度共生矩阵计算出星载SAR图像上建筑区与非建筑区的8种纹理特征统计量,根据巴氏距离进行特征选择,并通过主成分分析去除纹理特征之间的相关性,得到了最佳纹理特征分量,将所选的特征影像与原始图像进行波段组合,利用K均值聚类算法对组合后的图像进行非监督分类;最后通过对分类图像进行后处理并提取外部轮廓,提取了建筑区。以COSMO-SkyMed SAR影像为数据源进行了实验。结果表明该方法能够有效提取高分辨率星载SAR图像中的建筑区,提取效果明显优于未利用纹理特征的方法。     

基于对象建模的遥感影像建筑物提取方法

研究城镇建筑物的提取是遥感影像分析应用中的一项重要内容.遥感影像建筑物结构和光谱的多样性,使结构、光谱等特征的建筑物提取变得极其复杂.根据遥感影像的建筑物纹理区别于其它空间对象纹理的特点,为提高影像分辨率,提出Gabor纹理块的遥感影像对象模型方法应用于遥感影像城镇建筑物的提取.以整个城镇为对象,以建筑物、道路、绿地等不同城镇区域为组成对象的纹理块,建立基于纹理块的对象模型,利用模型进行遥感影像对象的纹理标定,最终提取出城镇建筑物.实验结果表明方法克服了建筑物结构复杂性和多样性以及背景环境的影响,能很好地从城镇遥感影像中提取建筑物.     

面向对象的损毁建筑物提取

地震灾害已经成为一种频发的自然灾害之一,在震后的灾害评估中,建筑物的倒塌情况是很重要的一项指标。本文用GeoEye影像提取海地地震中的损毁建筑物。由于海地地震时发生倒塌的房子大多处于比较老旧的地区,房屋比较密集,而且房顶结构也比较复杂,我们采取了一种基于规则集的方法通过分类将倒塌建筑物逐步与其他各种地物区分开来。即在影像多尺度分割的基础上,结合纹理特征及几何特征,通过规则集的方法构成分类树提取损毁房屋。另外在进行植被剔除时本文根据影像的特点提出了一种新的植被指数geo-NDVI,最后根据目视解译的结果对分类的结果进行了对比。     

融合类别语义特征的卷积神经网络建筑物提取

针对当前基于卷积神经网络的建筑提取方法在数据输入层对建筑语义信息利用不足的问题,提出一种融合类别语义特征的卷积神经网络建筑物提取方法。首先,从影像上提取形态学建筑指数,该指数能够直接表征建筑覆盖信息,是一种高层次的语义特征;然后,将该建筑语义特征作为原始影像的补充通道一起输入到卷积神经网络模型中训练,从数据层进一步增强建筑与背景的可分性。采用国际上公开的具有多种地物形态的标准建筑数据集验证本文方法的有效性。实验表明,所提出的方法取得了满意的精度(准确率为85.6%,召回率为93.1%,F指数为88.4%),相对于原始的RGB影像输入,建筑物语义特征的加入整体上提升了建筑物提取的精度。     

基于图像结构信息的城市房屋震害特征自动提取技术

由于常规边界检测算法难以完全满足城市房屋震害特征提取的需求，本研究提出了一种基于最优方法的边界跟踪方法，并利用ENVI IDL二次开发语言实现了基于区域结构和纹理统计特性相结合的损坏房屋自动识别算法。最后以1976年我国唐山地震的黑白航空影像为例，进行倒塌房屋的自动识别，得到了较为满意的结果。结果表明，利用本所提出的震害建筑自动识别方法，在提取震害房屋信息方面是有效的，其结果与目视结果大体一致。在完全倒塌的试验区，震害房屋的识别结果与目视解译的结果几乎完全一致，符合程度在99％左右；在其部分倒塌区，自动识别的结果与目视解译的结果符合程度也达到了89％。在这项自动识别理论中，由于采用了基于区域的图像处理与分析技术，所以保持了房屋建筑最基本的特性——区域特征，完全不同于基于光谱特征的震害房星识别技术。故而可以得到较好的自动识别结果。     

利用格式塔的高分辨率遥感影像建筑物提取

目的 格式塔心理学的理论基础为通过对事物的部分感知,实现对事物整体的认识。本文将该思想应用到建筑物提取中,提出一种兼顾目标细节及整体几何特征的高分辨率遥感影像建筑物提取方法。方法 首先,利用SIFT算法提取特征点作为候选边缘点;然后定义格式塔序列连续性原则判别边缘点,从而得到边缘点点集;并由边缘点点集拟合边缘,实现遥感影像建筑物提取。结果 利用提出算法,对WorldView-2遥感影像进行建筑物提取实验。通过与基于多尺度分割和区域合并的建筑物提取算法对比可以看出,提出算法能够更加准确、完整地提取出建筑物。采用分支因子、遗漏因子、检测率和完整性4个定量化指标对实验结果的定量评价,本文算法的检测率和完整性均大于对比算法,且本文算法的检测率均在95%以上,验证了提出基于格式塔理论的高分辨率遥感影像建筑物提取算法的有效性和准确性。结论 基于格式塔的高分辨率遥感影像建筑物提取算法能够准确刻画建筑物细节特征,同时兼顾建筑物整体几何轮廓,准确提取高分辨率遥感影像中的建筑物。本文算法针对高分辨率遥感影像,适用于提取边缘具有直线特征的建筑物。使用本文算法进行遥感影像建筑物提取时,提取精度会随分辨率降低而降低,建议实验影像分辨率在5 m以上。     

结合多种影像特征与CNN的城市建筑物提取

针对当前卷积神经网络对于城市建筑物纹理特征信息和多尺度信息利用的不足,提出了一种基于多种影像特征与卷积神经网络的城市建筑物提取方法,对结合尺度特征和纹理特征后的CNN模型的建筑物分类提取精度及其影响因素开展研究。方法采用局部二值模式来表达纹理特征,同时采用高斯金字塔提取多尺度特征,以此构建网络训练样本。基于此样本进行SegNet卷积网络训练,采用Softmax分类器完成建筑物粗提取,最后优化网络输出。研究表明,将纹理特征和尺度特征加入模型进行训练可以提高预测精度,其中精确率、召回率以及F1评分3个指标分别提升了8.01%、2.71%和4.98%。     

雷达遥感建筑物震害信息提取方法综述

合成孔径雷达遥感具有全天时、全天候工作的能力,在地震灾害应急中发挥了重要作用。回顾了雷达遥感建筑物震害信息提取技术的发展历史,总结了各种用于建筑物震害信息提取的雷达卫星的参数特征以及使用雷达遥感提取建筑物震害信息的震例。介绍了目前应用的各种雷达遥感建筑物震害信息提取方法,包括目视解译和计算机自动信息提取两种,其中后者包括基于纹理特征和极化特征的单时相方法、基于强度特征和相干特征的多时相方法。比较了这些方法的适用性和不足,并探讨了雷达遥感建筑物震害信息提取方法的发展趋势。     

A Summary of Research Methods for Seismic Damage Information Recognition of Polarized SAR Buildings

Synthetic Aperture Radar (SAR) can observe the Earth without the influence of the weather and sunlight, and Polarimetric SAR (PolSAR) even could acquire four kinds of polarization information at the same time. Therefore, extracting post-earthquake damage information by use of PolSAR has the advantage of timeliness and accuracy. This paper shows a summary of the methods for extracting seismic damage information based on PolSAR data. It firstly review the development of PolSAR and then summarizes the application and comparative analysis of the data types (multi-source data, multi-temporal data and single-temporal data) for extracting seismic damage of buildings in the past 10 years. Next, the methods of building earthquake damage extraction based on polarization decomposition and polarization characteristics and texture features is summarized. Finally, the research work is proposed to supplement the deficiency of PolSAR in earthquake damage extraction accuracy with the combination of geographic information data POI.     

Damage assessment in urban areas using post-earthquake airborne PolSAR imagery

Synthetic aperture radar (SAR) has often been used in earthquake damage assessment due to its extreme versatility and almost all-weather, day-and-night capability. In this article, we demonstrate the potential to use only post-event, high-resolution airborne polarimetric SAR (PolSAR) imagery to estimate the damage level at the block scale. Intact buildings with large orientation angles have a similar scattering mechanism to collapsed buildings; they are all volume-scattering dominant and reflection asymmetric, which seriously hampers the process of damage assessment. In this article, we propose a new damage assessment method combining polarimetric and spatial texture information to eliminate this deficiency. In the proposed method, the normalized circular-pol correlation coefficient is used first to identify intact buildings aligned parallel with the flight direction of the radar. The ‘homogeneity’ feature of the grey-level co-occurrence matrix (GLCM) is then introduced to distinguish building patches with large orientation angles from the severely damaged class. Furthermore, a new damage assessment index is also introduced to handle the assessment at the level of the block scale. To demonstrate the effectiveness of the proposed approach, the high-resolution airborne PolSAR imagery acquired after the earthquake that hit Yushu County, Qinghai Province of China, is investigated. By comparison with the damage validation map, the results confirm the validity of the proposed method and the advantage of further improving the assessment accuracy without external ancillary optical or SAR data.     

基于敏感特征向量的SAR图像灾害变化检测技术

为了解决阴雨云雾条件下光学遥感图像的应用局限性问题,针对典型的四类地表变化（堰塞湖、滑坡泥石流、部分倒塌建筑和严重倒塌建筑）分析SAR图像灰度和纹理特征的敏感程度,并提出敏感特征向量的概念;以综合利用了灰度差值和纹理差值的敏感特征向量作为评判因子,结合主成分分析技术和K均值聚类技术,提出了新的SAR图像灾害变化检测算法。该方法算法简单,检测效果好,并用两组ALOS SAR实验数据进行了证实。     

Extraction of earthquake-induced collapsed buildings using very high-resolution imagery and airborne lidar data

Extraction of urban building damage caused by earthquake disasters, from very-high-resolution (VHR) satellite imagery and related geospatial data, has been widely studied in the past decade. In this study, a multi-stage collapsed building detection method, using bi-temporal (pre- and post-earthquake) VHR images and post-earthquake airborne light detection and ranging (lidar) data, is proposed. Ground objects that are intact and significantly different from collapsed buildings, such as intact buildings, pavements, shadows, and vegetation, were first extracted using the post-event VHR image and lidar data and masked out. Collapsed buildings were then extracted by classifying the combined bi-temporal VHR images and texture images of the remaining area using a one-class classifier, the One-Class Support Vector Machine (OCSVM). A post-processing procedure was adopted to refine the obtained result. The proposed method was quantitatively evaluated and compared to two existing methods in Port au Prince, Haiti, which was heavily hit by an earthquake in January 2010. In the two comparative methods, data for the whole study area were directly used. In the first method, collapsed buildings were extracted directly using the OCSVM, while in the second method, buildings and pavements were removed from the extraction result of the first method. The results showed that the proposed method significantly outperformed the existing methods, with increases of 21% and 40%, respectively, in the kappa coefficient. The proposed method provides a fast and reliable method to detect collapsed urban buildings caused by earthquake disasters, and could also be applied to other study areas using similar data combinations.     

Detection of collapsed buildings caused by the 1999 Izmit,Turkey earthquake through digital analysis of post-event aerial photographs

In this study, the post-earthquake aerial photographs were digitally processed and analysed to detect collapsed buildings caused by the Izmit, Turkey earthquake of 17 August 1999. The selected area of study encloses part of the city of Golcuk, which is one of the urban areas most strongly hit by the earthquake. The analysis relies on the idea that if a building is collapsed, then it will not have corresponding shadows. The boundaries of the buildings were available and stored in a Geographical Information System (GIS) as vector polygons. The vector building polygons were used to match the shadow casting edges of the buildings with their corresponding shadows and to perform analyses in a building-specific manner. The shadow edges of the buildings were detected through a Prewitt edge detection algorithm. For each building, the agreement was then measured between the shadow producing edges of the building polygons and the thresholded edge image based on the percentage of shadow edge pixels. If the computed percentage value was below a preset threshold then the building being assessed was declared as collapsed. Of the 80 collapsed buildings, 74 were detected correctly, providing 92.50% producer's accuracy. The overall accuracy was computed as 96.15%. The results show that the detection of the collapsed buildings through digital analysis of post-earthquake aerial photographs based on shadow information is quite encouraging. It is also demonstrated that determining the optimum threshold value for separating the collapsed from uncollapsed buildings is important.     

Detecting damaged building parts in earthquake-damaged areas using differential seeded region growing

Acquiring information about earthquake-damaged buildings is essential for effective rescue and restoration operations. Building damage must be assessed to provide detailed information regarding the location and proportion of damage to individual buildings. Automatic processing of damage assessment is also critical in hastening relief efforts. Therefore, we propose a new method for automatically extracting damaged building parts and quantitatively assessing the damage to individual buildings caused by earthquakes. The proposed method consists of four parts: generating differential information, differential seeded region growing (DSRG), rule-based earthquake damage analysis, and accuracy assessment. First, differential information is automatically derived to extract the damage candidates. The damage candidates are then used as seed points for the region growing process to extract damaged building parts without requiring intervention by a human analyst. Then, designed automated extraction rules based on the condition of the collapsed or crushed buildings are used on the DSRG results. We applied the proposed method to both a residential area and a business area in Port-au-Prince, Haiti, and evaluated its accuracy using a visual comparison, a location-based assessment, and a proportion-based assessment. The results of the visual comparison were similar to the reference data, exhibiting location accuracies of 86% and 89% for the chosen residential and business areas, respectively. An assessment of the damage proportion to individual buildings was performed, which showed that the proposed method achieved accuracies of 81% and 84% for the residential and business areas, respectively, and was highly correlated with the reference data. The proposed method can accurately estimate damaged building parts, which can accelerate rapid relief actions in earthquake-damaged areas. In addition, the proposed method promotes cost-effective relief actions because it filters out many intact buildings without omitting damaged buildings.     

Detection of collapsed buildings with the aerial images captured from UAV基于低空无人机航拍图像的坍塌建筑物自动探测方法

In this paper, we present a method of detecting the collapsed buildings with the aerial images which are captured by an unmanned aerial vehicle (UAV) for the postseismic evaluation. Different from the conventional methods that apply the satellite images or the high-altitude UAV for the coarse disaster evaluation over large area, the purpose of this work is to achieve the accurate detection of collapsed buildings in small area from low altitude. By combining the motion and appearance features of collapsed buildings extracted from successive aerial images, each pixel in the input image will be measured by a statistical method where the background pixels will be penalized and the ones of collapsed buildings will be assigned with high value. The candidates of collapsed buildings will be established by integrating the extracted feature points into local groups with the online clustering algorithm. To reduce the false alarm caused by the complex background noise, each predicted candidate will be further verified by the temporal tracking framework where both the trajectory and the appearance of a candidate will be measured. The candidate of collapsed buildings that can survive through long time will be considered as true positive, otherwise rejected as a false alarm. Through extensive experiments, the efficiency and the effectiveness of proposed algorithm have been proved.