1.大连大学建筑工程学院,辽宁 大连 116622
2.大连大学辽宁省岩土与结构工程技术研究中心,辽宁 大连 116622
赵晓东(1969—),男,教授,博士。主要从事岩土工程、防灾减灾以及GIS工程模型耦合方面的研究。E-mail:xdong_zhao@126.com
收稿:2025-06-12,
修回:2025-08-25,
纸质出版:2026-10-28
移动端阅览
赵晓东,王曦阅.基于阈值修正的Inception‑v3深度学习模型及其地灾识别应用研究[J].防灾减灾工程学报,2026,46(05):962-971.
ZHAO Xiaodong,WANG Xiyue.Research on Inception‑v3 Deep Learning Model Based on Threshold Correction and Its Application to Geohazard Recognition[J].Journal of Disaster Prevention and Mitigation Engineering,2026,46(05):962-971.
赵晓东,王曦阅.基于阈值修正的Inception‑v3深度学习模型及其地灾识别应用研究[J].防灾减灾工程学报,2026,46(05):962-971. DOI: 10.13409/j.cnki.jdpme.20250612006.
ZHAO Xiaodong,WANG Xiyue.Research on Inception‑v3 Deep Learning Model Based on Threshold Correction and Its Application to Geohazard Recognition[J].Journal of Disaster Prevention and Mitigation Engineering,2026,46(05):962-971. DOI: 10.13409/j.cnki.jdpme.20250612006.
极端强降雨事件会诱发大量滑坡和崩塌灾害,造成大量人员伤亡和损失灾害。快速提取滑坡的特征及范围可为应急处理和灾害的风险防控提供有效的数据支撑。本研究利用灾后光学影像,基于改进的Inception‑v3深度学习模型进行智能识别,在纹理、形态等多层次特征提取下,通过地形的阈值修正判读结果,解决了云雾遮挡和浑水的不确定的判读问题。结果显示,在使用6°以下坡度阈值修正时,
F
值由0.76提升至0.896;测试集条件下,修正模型精确率为0.898,召回率为0.883,
F
值为0.890; 25 m×25 m网格划分(对应125×125像素)的分割图像在实验条件下平均识别耗时8分钟,较传统人工目视解译(6人×15小时/张)提升45倍,满足应急响应对时效性的要求,可为防灾减灾救灾提供快速、准确和有效的科学依据。
Extremely heavy rainfall events can induce a large number of landslides and collapse disasters
resulting in substantial casualties and losses. The rapid extraction of landslide characteristics and extent can provide effective data support for emergency response and disaster risk prevention and control. This study used post-disaster optical images for intelligent recognition based on an improved Inception-v3 deep learning model. Following multi-level feature extraction
including texture and morphology
the interpretation results were corrected using terrain thresholds
resolving interpretation uncertainties caused by cloud cover and muddy water. The results showed that the F value increased from 0.76 to 0.896 when using a slope threshold correction below 6°. On the test set
the corrected model achieved a precision of 0.898
a recall of 0.883
and an F value of 0.890. The segmented images divided into 25 m × 25 m grids (corresponding to 125 × 125 pixels) required an average of 8 minutes for recognition under experimental conditions
representing a 45-fold improvement in efficiency compared with traditional manual visual interpretation (6 people × 15 hours/image). The method meets the timeliness requirements of emergency response and can provide rapid
accurate
and effective scientific support for disaster prevention
mitigation
and relief.
张勤 , 赵超英 , 陈雪蓉 . 多源遥感地质灾害早期识别技术进展与发展趋势 [J]. 测绘学报 , 2022 , 51 ( 6 ): 885 - 896 .
Zhang Q , Zhao C Y , Chen X R . Technical progress and development trend of geological hazards early identification with multi-source remote sensing [J]. Acta Geodaetica et Cartographica Sinica , 2022 , 51 ( 6 ): 885 - 896 . (in Chinese)
刘传正 , 王建新 . 崩塌滑坡泥石流灾害链分类研究 [J]. 工程地质学报 , 2024 , 32 ( 5 ): 1573 - 1596 .
Liu C Z , Wang J X . Research on classification of collapse, landslide and debris flow disaster chains [J]. Journal of Engineering Geology , 2024 , 32 ( 5 ): 1573 - 1596 . (in Chinese)
张银胜 , 陈戈 , 段修贤 , 等 . 基于多层特征信息融合的滑坡图像分割模型 [J]. 测绘学报 , 2024 , 53 ( 11 ): 2201 - 2212 .
Zhang Y S , Chen G , Duan X X , et al . Landslide image segmentation model based on multi-layer feature information fusion [J]. Acta Geodaetica et Cartographica Sinica , 2024 , 53 ( 11 ): 2201 - 2212 . (in Chinese)
刘传正 , 米文忠 , 黄帅 . 论灾害事故预防应对的调查评估问题 [J]. 灾害学 , 2025 , 40 ( 1 ): 1 - 7, 85 .
Liu C Z , Mi W Z , Huang S . Research and evaluation issues in disaster accident prevention and response [J]. Journal of Catastrophology , 2025 , 40 ( 1 ): 1 - 7, 85 . (in Chinese)
许强 . 对地质灾害隐患早期识别相关问题的认识与思考 [J]. 武汉大学学报(信息科学版) , 2020 , 45 ( 11 ): 1651 - 1659 .
Xu Q . Understanding and consideration of related issues in early identification of potential geohazards [J]. Geomatics and Information Science of Wuhan University , 2020 , 45 ( 11 ): 1651 - 1659 . (in Chinese)
Prakash N , Manconi A , Loew S . Mapping landslides on EO data: Performance of deep learning models vs . traditional machine learning models [J]. Remote Sensing , 2020 , 12 ( 3 ): 346 .
Liu X R , Peng Y X , Lu Z L , et al . Feature-fusion segmentation network for landslide detection using high-resolution remote sensing images and digital elevation model data [J]. IEEE Transactions on Geoscience and Remote Sensing , 2023 , 61 : 4500314 .
姜万冬 , 席江波 , 李振洪 , 等 . 模拟困难样本的Mask R-CNN滑坡分割识别 [J]. 武汉大学学报(信息科学版) , 2023 , 48 ( 12 ): 1931 - 1942 .
Jiang W D , Xi J B , Li Z H , et al . Landslide detection and segmentation using mask R-CNN with simulated hard samples [J]. Geomatics and Information Science of Wuhan University , 2023 , 48 ( 12 ): 1931 - 1942 . (in Chinese)
张福浩 , 朱月月 , 赵习枝 , 等 . 地理因子支持下的滑坡隐患点空间分布特征及识别研究 [J]. 武汉大学学报(信息科学版) , 2020 , 45 ( 8 ): 1233 - 1244 .
Zhang F H , Zhu Y Y , Zhao X Z , et al . Spatial distribution and identification of hidden danger points of landslides based on geographical factors [J]. Geomatics and Information Science of Wuhan University , 2020 , 45 ( 8 ): 1233 - 1244 . (in Chinese)
Cao Y , Yin K L , Zhou C , et al . Establishment of landslide groundwater level prediction model based on GA-SVM and influencing factor analysis [J]. Sensors , 2020 , 20 ( 3 ): 845 .
Wang H J , Zhang L M , Yin K S , et al . Landslide identification using machine learning [J]. Geoscience Frontiers , 2021 , 12 ( 1 ): 351 - 364 .
Zhang K X , Wu X L , Niu R Q , et al . The assessment of landslide susceptibility mapping using random forest and decision tree methods in the Three Gorges Reservoir area, China [J]. Environmental Earth Sciences , 2017 , 76 ( 11 ): 405 .
赵通 , 张双成 , 何晓宁 , 等 . 改进的DeepLabV 3+ 模型用于震后高分遥感影像滑坡识别 [J]. 遥感学报 , 2024 , 28 ( 9 ): 2293 - 2305 .
Zhao T , Zhang S C , He X N , et al . Improved DeepLabV 3+ model for landslide identification in high-resolution remote sensing images after earthquakes [J]. National Remote Sensing Bulletin , 2024 , 28 ( 9 ): 2293 - 2305 . (in Chinese)
Garcia G P B , Soares L P , Espadoto M , et al . Relict landslide detection using deep-learning architectures for image segmentation in rainforest areas: A new framework [J]. International Journal of Remote Sensing , 2023 , 44 ( 7 ): 2168 - 2195 .
Shi W Z , Zhang M , Ke H F , et al . Landslide recognition by deep convolutional neural network and change detection [J]. IEEE Transactions on Geoscience and Remote Sensing , 2021 , 59 ( 6 ): 4654 - 4672 .
Liu P , Wei Y M , Wang Q J , et al . Research on post-earthquake landslide extraction algorithm based on improved U-Net model [J]. Remote Sensing , 2020 , 12 ( 5 ): 894 .
王家柱 , 铁永波 , 白永健 , 等 . 机器学习在斜坡地质灾害领域的应用现状与展望 [J]. 水文地质工程地质 , 2025 , 52 ( 4 ): 228 - 244 .
Wang J Z , Tie Y B , Bai Y J , et al . Application and prospects of machine learning for rockfalls, landslides and debris flows [J]. Hydrogeology & Engineering Geology , 2025 , 52 ( 4 ): 228 - 244 . (in Chinese)
林秋怡 , 左仁广 . 迁移学习及其在固体地球科学中的应用 [J]. 地质科技通报 , 2025 , 44 ( 1 ): 346 - 356 .
Lin Q Y , Zuo R G . Transfer learning and its application in solid earth geoscience [J]. Bulletin of Geological Science and Technology , 2025 , 44 ( 1 ): 346 - 356 . (in Chinese)
Ma H J , Liu Y L , Ren Y H , et al . Improved CNN classification method for groups of buildings damaged by earthquake, based on high resolution remote sensing images [J]. Remote Sensing , 2020 , 12 ( 2 ): 260 .
Manzo C , Valentini E , Taramelli A , et al . Spectralcharacterization of coastalsedimentsusing field spectral libraries, airborne hyperspectral images and topographic LiDAR data (FHyL) [J]. International Journal of Applied Earth Observation and Geoinformation , 2015 , 36 : 54 - 68 .
Hugenholtz C H , Wolfe S A . Rates and environmental controls of aeolian dust accumulation, Athabasca River Valley, Canadian Rocky Mountains [J]. Geomorphology , 2010 , 121 ( 3/4 ): 274 - 282 .
Szegedy C , Vanhoucke V , Ioffe S , et al . Rethinking the inception architecture for computer vision [C]∥ 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) . New York : IEEE Press , 2016 : 2818 - 2826 .
陶亮亮 , 王雨琦 . 基于雷达和光学遥感数据的云污染区域光谱重建算法 [J]. 农业工程学报 , 2024 , 40 ( 17 ): 216 - 223 .
Tao L L , Wang Y Q . Reconstructing spectral information in cloud contaminated regions using object-oriented approach [J]. Transactions of the Chinese Society of Agricultural Engineering , 2024 , 40 ( 17 ): 216 - 223 . (in Chinese)
牛朝阳 , 高欧阳 , 刘伟 , 等 . 光学遥感图像滑坡检测研究进展 [J]. 航天返回与遥感 , 2023 , 44 ( 3 ): 133 - 144 .
Niu C Y , Gao O Y , Liu W , et al . Research progress of landslide detection in optical remote sensing images [J]. Spacecraft Recovery & Remote Sensing , 2023 , 44 ( 3 ): 133 - 144 . (in Chinese)
赵晓东 , 杨华 , 王曦阅 . 三维点云数据微地形特征量的提取及应用研究 [J]. 测绘科学 , 2024 , 49 ( 5 ): 133 - 142 .
Zhao X D , Yang H , Wang X Y . Study on extraction of micro-terrain features from 3D point clouds and its application [J]. Science of Surveying and Mapping , 2024 , 49 ( 5 ): 133 - 142 . (in Chinese)
日本国土交通省国土地理院 . 平成30年(2018年)北海道胆振東部地震に関する情報 [EB/OL].( 2018-10-09 )[ 2025-06-12 ]. https:∥www.gsi.go.jp/BOUSAI/H30-hokkaidoiburi-east-earthquake-index.html https://www.gsi.go.jp/BOUSAI/H30-hokkaidoiburi-east-earthquake-index.html .
0
浏览量
9
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
苏公网安备32010202012147号
