结合空间域和频率域纹理特征的深度学习建筑震害提取方法--以2023土耳其地震为例  

A Deep Learning Approach for Earthquake Damage Extraction in Buildings by Integrating Spatial and Frequency Domain Texture Features-A Case Study of the 2023 Turkey Earthquake

作　　者：朱贵钰 翟玮 杜建清 ZHU Guiyu;ZHAI Wei;DU Jianqing(Lanzhou Institute of Seismology,CEA,Lanzhou 730000,China;Key Laboratory of Loess Earthquake Engineering of China Earthquake Administration,Lanzhou 730000,China;Lanzhou Institute of Geotechnique and Earthquake,China Earthquake Administration,Lanzhou 730000,China)

机构地区：[1]中国地震局兰州地震研究所,甘肃兰州730000 [2]中国地震局黄土地震工程重点实验室,甘肃兰州730000 [3]中国地震局兰州岩土工程与地震研究所,甘肃兰州730000

出　　处：《遥感技术与应用》2025年第1期89-97,共9页Remote Sensing Technology and Application

基　　金：国家自然科学基金项目(42371404、41601479);甘肃省科技计划项目自然科学基金面上项目(22JR5RA822);甘肃省重点人才项目(11276679015);甘肃省地震局创新团队专项(2019TD⁃01⁃02);中国地震局地震预测研究所基本科研业务费专项(2021IESLZ4)。

摘　　要：大型地震灾害发生时都会造成大量财产损失甚至人员伤亡,因此,灾后第一时间对灾情作出判断极为重要。合成孔径雷达(Synthetic Aperture Radar,SAR)具有全天时、全天候,不易受光照和气候条件影响的优势,使用SAR影像进行变化检测在震后救援与损失估计、洪水波及范围检测、城镇化研究以及海岸线提取等多个领域备受关注。为此,提出一种结合空间域和频率域纹理特征的深度学习建筑震害提取方法,能够较好地识别倒塌建筑物与未倒塌建筑。以2023年2月6日土耳其7.8级地震的卡赫拉曼马拉什(Kahramanmaras)区域为例,该区域在此次地震中损毁最严重。本研究在深度学习网络进行分类的过程中加入了空间域特征和频率域特征。经过大量实验计算,该方法的分类精度达80.98%,远高于原始图像的分类精度47.84%。同时,此结果也高于仅使用空间域特征的分类精度73.30%,以及仅使用频率域特征的分类精度73.42%。该方法能够为地震灾后的受灾情况及灾害评估提供基础支持。Large-scale Earthquake disasters often result in significant losses and even casualties.Making prompt assessments of the disaster situation is crucial in the aftermath.Synthetic Aperture Radar(SAR)possesses ad⁃vantages such as all-weather and all-day capabilities,as well as resilience to lighting and weather conditions.Therefore,the use of SAR imagery for change detection has garnered significant attention in various fields,in⁃cluding post-earthquake rescue and damage estimation,flood extent detection,urbanization studies,and coast⁃line extraction.In this context,this paper proposes a deep learning-based earthquake damage extraction method that integrates spatial and frequency domain texture features.The method demonstrates a robust capability to identify collapsed and intact structures.Using the 2023 earthquake in Kahramanmaras,Turkey,as a case study,the region severely affected by the earthquake,this research incorporates both spatial and frequency do⁃main features into the deep learning network for classification.Experimental results show that the proposed method achieves a classification accuracy of 80.98%,significantly surpassing the original image's accuracy of 47.84%.Moreover,this accuracy is higher than using only spatial domain features(73.30%)or only frequency domain features(73.42%).The proposed method in this study provides fundamental support for post-earth⁃quake disaster assessment and situational awareness.

关 键 词：SAR 建筑物震害 深度学习 空间域纹理特征 频率域纹理特征 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] P237[自动化与计算机技术—控制科学与工程] P315.9[天文地球—摄影测量与遥感]