Classification of Damaged Grade on Rural Houses after Flood Disaster Based on Deep Convolutional Neural Network and Transfer Learning
-
摘要: 洪涝灾害对房屋等建筑物会造成巨大损害，灾后房屋破坏等级鉴定对保障人民生命安全至关重要，而传统的人工鉴定方法，消耗较多的人力、财力及时间等资源.为此，基于河南郑州“7·20”特大暴雨引发的农村房屋破坏数据，采用深度卷积神经网络（CNN）理论，得到灾后房屋危险等级智能分类模型.首先采用AlexNet、VGGNet、GoogleNet和ResNet四种经典的深度CNN架构，对数据集进行训练、验证和测试，得到4种灾后房屋危险等级智能分类模型，然后结合迁移学习方法训练CNN提高模型的泛化能力，并选择效果较优的ResNet-50为分类主模型，最后分析CNN架构中超参数的影响.结果表明：ResNet-50在学习率为0.000 5，epoch为50，batch_size为16时网络训练结果最优，其测试集的预测准确率达到了95.5%；此外，房屋危险等级特征的可视化分析明确了模型分类的机理及准确性.试验表明基于迁移学习的识别模型准确率较高，为农村房屋洪涝灾害后受损等级分类模型提供参考.Abstract: Flood disasters can cause great damage to buildings, and identification of post-disaster house damage grade is very important to ensure people's safety. However, the traditional method of artificial identification costs a lot of manpower, financial resources, and time. Based on the data of rural housing damage caused by the "7.20" heavy rainstorm in Zhengzhou, Henan Province, the deep Convolutional neural network (CNN) theory was used to obtain the intelligent classification model of post-disaster house damage grade. Four classic deep CNN architectures, AlexNet, VGGNet, GoogleNet, and ResNet, were used to train, validate and test the data sets, and four intelligent classification models of post-disaster house damage grade were obtained. The weights of the pre-training model were fine-tuned to improve the generalization of the model, then the ResNet-50 with the transfer learning was selected as the main model of the classification. Finally, the influence of hyperparameters in CNN architecture was analyzed. The results show that when the learning rate of the ResNet network was 0.000 5, the epoch was 50, and the batch_size was 16, the network training result was optimal, and the prediction accuracy of the test set reached 95.5%; the visual analysis of the characteristics of the housing risk level clarified the mechanism and accuracy of the model classification. Experimental results show that the ensemble model had a high accuracy rate, which provided an idea and explored an approach for classification of damaged grade rural houses after flood disaster.
-
图 1 河南郑州“7·20”特大暴雨灾害引发的部分洪涝灾区
Fig. 1. Some flood disaster areas caused by Henan Zhengzhou "7.20" rainstorm
图 5 基于迁移学习的灾后房屋图像识别方法流程
Fig. 5. Flow chart of post-disaster housing image recognition method based on transfer learning
图 6 不同学习率结果对比（ResNet训练分类模型）
Fig. 6. Result comparison diagram of different learning rates(ResNet training classification model)
图 10 四种分类模型的预测结果混淆矩阵
Fig. 10. Confusion matrix of prediction results of four classification models
图 11 基于Grad-CAM方法的灾后房屋特征评估结果可视化
Fig. 11. Visualization of post-disaster housing feature assessment results based on Grad-CAM
图 12 迁移学习前后的灾后房屋特征评估结果可视化
Fig. 12. Visualization of post-disaster housing feature assessment results before and after transfer learning
表 1 ResNet模型的准确率、精确率和召回率
Table 1. Accuracy, precision and recall of ResNet model
数据集 类型 图像数量 TP TN FP FN 准确率 精确率 召回率 测试集 a 140 132 421 3 8 98.05 97.8 94.3 b 140 132 406 18 8 95.39 88.0 94.3 c 144 137 413 7 7 97.52 95.1 95.1 d 140 135 424 0 5 99.11 1.00 96.4 
下载: 导出CSV
表 2 不同CNN架构迁移学习前后结果对比
Table 2. Result comparison before and after transfer learning of different CNN architectures
模型 迁移学习 训练集准确率 验证集准确率 测试集准确率 测试时长(ms) AlexNet 否 91.1 87.7 64.4 11.6 是 93.8 90.5 84 11.7 GoogLeNet 否 92.9 82.4 69.8 10.2 是 94.9 90.1 78.9 10.5 VGGNet-16 否 85.3 86.7 73.3 11.1 是 92.8 93.7 88.3 12.4 ResNet-50 否 87 84.7 78.9 9.5 是 95.6 97.6 95.5 9.9 ResNet-101 否 88.4 90.6 77.7 11.2 是 95.3 96.5 93.3 11.2
下载: 导出CSV
-
Ayenu-Prah, A., Attoh-Okine, N., 2008. Evaluating Pavement Cracks with Bidimensional Empirical Mode Decomposition. Eurasip Journal on Advances in Signal Processing, 2008: 1-7. https://doi.org/10.1155/2008/861701 Cha, Y. J., Choi, W., Buyukozturk, O., 2017. Deep Learning-Based Crack Damage Detection Using Convolutional Neural Networks. Computer-Aided Civil and Infrastructure Engineering, 32: 361-378. https://doi.org/10.1111/mice.12263 Chen, J. Y., Zhou, M. L., Zhang, D. M., et al., 2021. Quantification of Water Inflow in Rock Tunnel Faces via Convolutional Neural Network Approach. Automation in Construction, 123: 103526. https://doi.org/10.1016/j.autcon.2020.103526 Chen, Z. Y., Kong, F., 2022. Study on Fragmentation of Emergency Management during"7·20"Extreme Rainstorm Flood Disaster in Zhengzhou of Henan Province and Relevant Comprehensive Treatment. Water Resources and Hydropower Engineering, 53(8): 1-14 (in Chinese with English abstract). Cheng, H. D., Shi, X. J., Glazier, C., 2003. Real-Time Image Thresholding Based on Sample Space Reduction and Interpolation Approach. Journal of Computing in Civil Engineering, 17(4): 264-272. https://doi.org/10.1061/(ASCE)0887-3801(2003)17:4(264) Feng, C., Pan, J. G., Li, C., et al., 2022. Fault High-Resolution Recognition Method Based on Deep Neural Network. Earth Science, 1-15 (in Chinese with English abstract). Gao, Y., Mosalam, K. M., 2018. Deep Transfer Learning for Image-Based Structural Damage Recognition. Computer-Aided Civil and Infrastructure Engineering, 33(9): 748-768. https://doi.org/10.1111/mice.12363 He, K., Zhang, X., Ren, S., et al., 2016. Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016: 770-778. https://doi.org/10.1109/CVPR.2016.90 Huang, F. M., Hu, S. Y., Yan, X. Y., et al., 2022a. Landslide Susceptibility Prediction and Identification of Its Main Environmental Factors Based on Machine Learning Models. Bulletin of Geological Science and Technology, 41(2): 79-90 (in Chinese with English abstract). Huang, F. M., Li, J. F., Wang, J. Y., et al., 2022b. Modelling Rules of Landslide Susceptibility Prediction Considering the Suitability of Linear Environmental Factors and Different Machine Learning Models. Bulletin of Geological Science and Technology, 41(2): 44-59 (in Chinese with English abstract). Huang, Y. X., Xu, B. G., 2006. Automatic Inspection of Pavement Cracking Distress. Journal of Electronic Imaging, 15(1): 1-6. https://doi.org/10.1117/1.2177650 Jahanshahi, M. R., Jazizadeh, F., Masri, S. F., et al., 2013. Unsupervised Approach for Autonomous Pavement-Defect Detection and Quantification Using an Inexpensive Depth Sensor. Journal of Computing in Civil Engineering, 27(6): 743-754. https://doi.org/10.1061/(ASCE)CP.1943-5487.0000245 Kirschke, K. R., Velinsky, S. A., 1992. Histogram-Based Approach for Automated Pavement-Crack Sensing. Journal of Transportation Engineering, 118(5): 700-710. https://doi.org/10.1061/(ASCE)0733-947X(1992)118:5(700) Kohavi, R., Provost, F., 1998. Confusion Matrix. Machine Learning, 30(2-3): 271-274. Kong, F., 2020. System and Capacity Building of Disaster Prevention, Mitigation and Relief in Rural China: Significance, Current Situation, Challenges and Countermeasures. Disaster Reduction in China, 21: 10-13 (in Chinese). Krizhevsky, A., Sutskever, I., Hinton, G., 2012. ImageNet Classification with Deep Convolutional Neural Networks. Communications of the ACM, 60: 84-90. https://doi.org/10.1145/3065386 Le Cun, Y., Jackel, L. D., Boser, B., et al., 1989. Handwritten Digit Recognition: Applications of Neural Network Chips and Automatic Learning. IEEE Communications Magazine, 27(11): 41-46. https://doi.org/10.1109/35.41400 Li, S., Chen, J., Liu, C., et al., 2021. Mineral Prospectivity Prediction via Convolutional Neural Networks Based on Geological Big Data. Journal of Earth Science, 32(2): 327-347. https://doi.org/10.1007/s12583-020-1365-z Li, Z. B., Li, M., Zhao, Y. Y., et al., 2021. Iced Pomfret Freshness Evaluation Method Based on Improved VGG-19 Convolutional Neural Networks. Transactions of the Chinese Society of Agricultural Engineering, 37(22): 286-294 (in Chinese with English abstract). doi: 10.11975/j.issn.1002-6819.2021.22.033 Liu, B., Xu, H. W., Li, C. Z., et al., 2022. Apple Leaf Disease Identification Method Based on Snapshot Ensemble CNN. Transactions of the Chinese Society for Agricultural Machinery, 53(6): 286-294 (in Chinese with English abstract). Liu, H. L., Zhang, R. H., Liu, D. S., et al., 2022. Study on Characteristics of Physical and Mechanical Parameters of Engineering Geology Based on Data Fusion Technique. Journal of Civil and Environmental Engineering, 44(6): 1-11 (in Chinese with English abstract). Makantasis, K., Protopapadakis, E., Doulamis, A., et al., 2015. Deep Convolutional Neural Networks for Efficient Vision Based Tunnel Inspection. 2015 IEEE International Conference on Intelligent Computer Communication and Processing (ICCP), Cluj-Napoca, 335-342. Nisanth, A., Mathew, A., 2014. Automated Visual Inspection of Pavement Crack Detection and Characterization. International Journal of Technology and Engineering System, 6(1): 14-20. http://pdfs.semanticscholar.org/4ca3/af17f968dbf22c9ed161c5a8b9efa9e01fd7.pdf Oliveira, H., Correia, P. L., 2009. Automatic Road Crack Segmentation Using Entropy and Image Dynamic Thresholding. 17th European Signal Processing Conference, Glasgow, 622-626. Our Correspondent, 2022. The "July 20" Rainstorm Disaster in Zhengzhou, Henan Reshaped the Concept of Urban Construction. Chinese Emergency Management, (2) : 6-7 (in Chinese with English abstract). doi: 10.1007/s11069-022-05792-z Ouyang, W., Xu, B., 2013. Pavement Cracking Measurements Using 3D Laser-Scan Images. Measurement Science and Technology, 24(10): 105204. https://doi.org/10.1088/0957-0233/24/10/105204 Santhi, B., Krishnamurthy, G., Siddharth, S., et al., 2012. Automatic Detection of Cracks in Pavements Using Edge Detection Operator. Journal of Theoretical and Applied Information Technology, 36(2): 199-205. http://www.jatit.org/volumes/Vol36No2/6Vol36No2.pdf Simonyan, K., Zisserman, A., 2014. Very Deep Convolutional Networks for Large-Scale Image Recognition. Computer Science, 14124313. https://doi.org/10.48550/arXiv.1409.1556 Süzen, A., Alkan Çakiroğlu, M., 2020. Assessment and Application of Deep Learning Algorithms in Civil Engineering. Journal of Science and Engineering, 7(3): 906-922. https://doi.org/10.31202/ECJSE.679113 Szegedy, C., Liu, W., Jia, Y., et al., 2015. Going Deeper with Convolutions. 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, 1-9. https://doi.org/10.1109/CVPR.2015.7298594 Thenmozhi, K., Reddy, U. S., 2019. Crop Pest Classification Based on Deep Convolutional Neural Network and Transfer Learning. Computers and Electronics in Agriculture, 164(C): 104906. Wang, K. C. P., 2011. Elements of Automated Survey of Pavements and a 3D Methodology. Journal of Modern Transportation, 19(1): 51-57. https://doi.org/10.1007/BF03325740 Yao, M., Li, X., Yuan, J. D., et al., 2023. Deep Learning Characterization Method of Rock Mass Conditions based on TBM Rock Breaking Data. Earth Science, 48(5): 1908-1922 (in Chinese with English abstract). Ying, L., Salari, E., 2010. Beamlet Transform-Based Technique for Pavement Crack Detection and Classification. Computer-Aided Civil and Infrastructure Engineering, 25(8): 572-580. https://doi.org/10.1111/j.1467-8667.2010.00674.x Zeiler, M. D., Fergus, R., 2014. Visualizing and Understanding Convolutional Networks. In: Proceedings of the European Conference on Computer Vision. Springer, Cham, 818-833. https://doi.org/10.1007/978-3-319-10590-1_53 Zhang, A., Wang, K., Li, B., et al., 2017a. Automated Pixel-Level Pavement Crack Detection on 3D Asphalt Surfaces Using a Deep-Learning Network. Computer-Aided Civil and Infrastructure Engineering, 32(10): 805-819. https://doi.org/10.1111/mice.12297 Zhang, A., Wang, K. C. P., Ai, C. F., 2017b. 3D Shadow Modeling for Detection of Descended Patterns on 3D Pavement Surface. Journal of Computing in Civil Engineering, 31(4): 1-13. https://doi.org/10.1061/(ASCE)CP.1943-5487.0000661 Zhang, A., Wang, K. C. P., 2017. The Fast Prefix Coding Algorithm (FPCA) for 3D Pavement Surface Data Compression. Computer-Aided Civil and Infrastructure Engineering, 32(3): 173-190. https://doi.org/10.1111/mice.12243 Zhang, J. Y., Wang, Y. T., He, R. M., et al., 2016. Discussion on the Urban Flood and Waterlogging and Causes Analysis in China. Advances in Water Science, 27(4): 485-491 (in Chinese with English abstract). http://en.cnki.com.cn/Article_en/CJFDTOTAL-SKXJ201604001.htm Zhang, W. G., He, Y. W., Wang, L. Q., et al., 2023. Machine Learning Solution for Landslide Susceptibility Based on Hydrographic Division: Case Study of Fengjie County in Chongqing. Earth Science, 48(5): 2024-2038 (in Chinese with English abstract). Zhang, W. G., Li, H. R., Wu, C. Z., et al., 2021a. Stability Assessment of Underground Entry-Type Excavations Using Data-Driven RF and KNN Methods. Journal of Hunan University (Natural Sciences), 48(3): 164-172 (in Chinese with English abstract). Zhang, W. G., Tang, L. B., Chen, F. Y., et al., 2021b. Prediction for TBM Penetration Rate Using Four Hyperparameter Optimization Methods and Random Forest Model. Journal of Basic Science and Engineering, 29(5): 1186-1200 (in Chinese with English abstract). doi: 10.1007/978-981-16-6835-7_8 本刊记者, 2022. 河南郑州"7·20"特大暴雨灾害重塑城市建设理念尊重自然系统谋划立足当前着眼长远. 中国应急管理, (2): 6-7. https://www.cnki.com.cn/Article/CJFDTOTAL-YIGU202202006.htm 谌舟颖, 孔锋, 2022. 河南郑州"7·20"特大暴雨洪涝灾害应急管理碎片化及综合治理研究. 水利水电技术(中英文), 53(8): 1-14. https://www.cnki.com.cn/Article/CJFDTOTAL-SJWJ202208001.htm 丰超, 潘建国, 李闯, 等, 2022. 基于深度神经网络的断层高分辨率识别方法. 地球科学, 1-15. 黄发明, 胡松雁, 闫学涯, 等, 2022a. 基于机器学习的滑坡易发性预测建模及其主控因子识别. 地质科技通报, 41(2): 79-90. https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202202008.htm 黄发明, 李金凤, 王俊宇, 等, 2022b. 考虑线状环境因子适宜性和不同机器学习模型的滑坡易发性预测建模规律. 地质科技通报, 41(2): 44-59. https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202202005.htm 孔锋, 2020. 我国农村防灾减灾救灾体系和能力建设: 意义、现状、挑战和对策. 中国减灾, 21: 10-13. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGJI202021009.htm 李振波, 李萌, 赵远洋, 等, 2021. 基于改进VGG-19卷积神经网络的冰鲜鲳鱼新鲜度评估方法. 农业工程学报, 37(22): 286-294. https://www.cnki.com.cn/Article/CJFDTOTAL-NYGU202122033.htm 刘斌, 徐皓玮, 李承泽, 等, 2022. 基于快照集成卷积神经网络的苹果叶部病害程度识别. 农业机械学报, 53(6): 286-294. https://www.cnki.com.cn/Article/CJFDTOTAL-NYJX202206030.htm 刘汉龙, 章润红, 刘东升, 等, 2022. 基于数据融合的工程地质物理力学参数特征研究. 土木与环境工程学报(中英文), 44(6): 1-11. https://www.cnki.com.cn/Article/CJFDTOTAL-JIAN202206001.htm 姚敏, 李旭, 原继东, 等, 2023. 基于TBM破岩数据的岩体条件深度学习表征方法. 地球科学, 48(5): 1908-1922. https://www.cnki.com.cn/Article/CJFDTOTAL-DQKX202305016.htm 张建云, 王银堂, 贺瑞敏, 等, 2016. 中国城市洪涝问题及成因分析. 水科学进展, 27(4): 485-491. https://www.cnki.com.cn/Article/CJFDTOTAL-SKXJ201604001.htm 仉文岗, 何昱苇, 王鲁琦, 等, 2023. 基于水系分区的滑坡易发性机器学习分析方法——以重庆市奉节县为例. 地球科学, 48(5): 2024-2038. https://www.cnki.com.cn/Article/CJFDTOTAL-DQKX202305028.htm 仉文岗, 李红蕊, 巫崇智, 等, 2021a. 基于RF和KNN的地下采场开挖稳定性评估. 湖南大学学报(自然科学版), 48(3): 164-172. https://www.cnki.com.cn/Article/CJFDTOTAL-HNDX202103017.htm 仉文岗, 唐理斌, 陈福勇, 等, 2021b. 基于4种超参数优化算法及随机森林模型预测TBM掘进速度. 应用基础与工程科学学报, 29(5): 1186-1200. https://www.cnki.com.cn/Article/CJFDTOTAL-YJGX202105009.htm -
点击查看大图
图(12) / 表(2)
计量
- 文章访问数: 863
- HTML全文浏览量: 549
- PDF下载量: 88
- 被引次数: 0
