[1]陈晓艺,陆一鸣,沈加炜,等.基于深度学习的灾后建筑物损坏程度检测综述[J].计算机技术与发展,2023,33(09):1-7.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 001]
　CHEN Xiao-yi,LU Yi-ming,SHEN Jia-wei,et al.Review of Post-disaster Building Damage Detection Based on Deep Learning[J].,2023,33(09):1-7.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 001]

点击复制

基于深度学习的灾后建筑物损坏程度检测综述()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年09期

页码:

1-7

栏目:

综述

出版日期:

2023-09-10

文章信息/Info

Title:

Review of Post-disaster Building Damage Detection Based on Deep Learning

文章编号:

1673-629X(2023)09-0001-07

作者:

陈晓艺¹ ; 陆一鸣² ; 沈加炜¹ ; 钱美玲¹ ; 陆卫忠¹; 3; 4

1. 苏州科技大学 电子与信息工程学院,江苏 苏州 215009;

2. 苏州科技大学天平学院,江苏 苏州 215009;

3. 苏州科技大学 苏州智慧城市研究院,江苏 苏州 215009;

4. 苏州科技大学 苏州市虚拟现实智能交互及应用技术重点实验室,江苏 苏州 215009

Author(s):

CHEN Xiao-yi1 ; LU Yi-ming2 ; SHEN Jia-wei1 ; QIAN Mei-ling1 ; LU Wei-zhong1; 3; 4

1. School of Electronic and Information Engineering,Suzhou University of Science and Technology,Suzhou 215009,China;

2. Suzhou University of Science and Technology Tianping College,Suzhou 215009,China;

3. Suzhou Institute of Smart City,Suzhou University of Science and Technology,Suzhou 215009,China;

4. Suzhou Key Laboratory of Virtual Reality Intelligent Interaction and Application Technology,Suzhou University of Science and Technology,Suzhou 215009,China

关键词:

遥感图像; 深度学习; 计算机视觉; 自然灾害; 建筑物损坏程度检测

Keywords:

remote sensing image; deep learning; computer vision; natural disasters; building damage detection

分类号:

TP305

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 09. 001

摘要:

遥感图像分类和语义分割是一项非常有应用价值的计算机视觉任务。 由于现实生活对遥感信息有更高的需求,使计算机视觉领域中高分辨率遥感图像研究日益活跃。?

其广泛应用于国土资源监测、道路提取和土地划分等领域。 自然灾害后建筑物损害程度检测也作为其应用领域之一,目的是对灾后建筑物损坏程度进行相关检测和评估。 近年来,随着深度学习的发展,遥感图像领域取得巨大进展,深度学习在遥感图像分类和语义分割领域中的应用获得了巨大的成功,使其解析遥感图像信息和提取底物特征的速度更快,也在很大程度上提高了处理遥感图像相关任务的准确性。 因此,深度学习中的计算机视觉技术对自然灾害后建筑物损害程度检测具有很大帮助。 该文介绍了基于深度学习的自然灾害后建筑物损坏程度检测的相关任务、难点和发展现状。 接着对 xBD 数据集进行介绍,并说明了不同算法模型的相关评价标准。然后对深度学习方法中几种应用于建筑物损坏程度检测的卷积神经网络模型进行总结和对比。 最后对其存在的问题及未来可能的发展方向进行了讨论。

Abstract:

Remote sensing image classification and semantic segmentation is a very valuable computer vision task. Due to the higherdemand for remote sensing information?

in real life,the research on high-resolution remote sensing images in the field of computer visionis increasingly active. It is widely used in the fields of land resource monitoring,road extraction and land division. The detection of thedamage degree of buildings after natural disasters is also one of its application fields,and the purpose is to detect and evaluate the damagedegree of buildings after disasters. In recent years,with the development of deep learning,great progress has been made in the field ofremote sensing images. The application of deep learning in remote sensing image classification and semantic segmentation has achievedgreat success,making it faster to analyze remote sensing image information and extract substrate features. It also greatly improves theaccuracy of processing remote sensing image-related tasks. Therefore,the computer vision technology in deep learning is of great help tothe detection of the damage degree of buildings after natural disasters. We introduce the related tasks,difficulties and development statusof building damage degree detection after natural disasters based on deep learning. Then,the xBD dataset is introduced,and the related evaluation criteria of different algorithm models are explained. Then, several convolutional neural network models applied to buildingdamage detection in deep learning methods are summarized and compared. Finally,the existing problems and possible future developmentdirections are discussed.

常用功能