Computer vision-based two-step inspection method for spalling and cracks of building facades

被引：0
|
作者：
Xia Z. [1 ,2 ]
Ma L. [1 ,2 ]
Shan J. [1 ,2 ]
Lu X. [1 ,2 ]
机构：
[1] Department of Disaster Mitigation for Structures, Tongji University, Shanghai
[2] State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai
来源：
Jianzhu Jiegou Xuebao/Journal of Building Structures | 2023年 / 44卷 / 02期
关键词：
building facades; computer vision; cracks detection; damage evaluation; digital image processing; spalling detection;
D O I：
10.14006/j.jzjgxb.2022.B069
中图分类号：
学科分类号：
摘要：
In the earthquake disasters, rapid inspection of building facades damage can contribute to disaster assessment and reconstruction. To realize automated non-contact detection of spalling and cracks on building facades, a two-step inspection method combing deep learning and image processing for fa ades damage localization and quantitative assessment is proposed. By establishing and labeling a dataset of earthquake-damaged buildings in the old county of Beichuan, an object detection network based on the Focal Loss function model is used to locate the bounding-box in breakage areas and overcome the unbalanced problem of positive and negative sample in small-scale area. Based on the extracted single image of breakage areas, a processing flow combing image filtering, threshold segmentation and morphological manipulation is proposed for pixel-level segmentation and extraction of spalling and cracks on tile wall. Further, the mapping relationship between pixel and physical scale is combined to complete the calculation and identification of key breakage information such as spalling area, cracks width. Results show that the method can perform better localization and extraction of wall spalling and cracks on post-earthquake building facades, achieve pixel-level segmentation of wall damage and accurately identify key physical parameters, and achieve accurate and rapid detection of damage on post-earthquake buildings. © 2023 Science Press. All rights reserved.
引用
收藏
页码：207 / 216
页数：9
相关论文
共 17 条
  • [1] SUN Baitao, ZHANG Lei, Comparison of post-earthquake emergency assessment researches on building damage and safety in China and overseas, China Civil Engineering Journal, 50, 7, pp. 8-19, (2017)
  • [2] PANERU S, JEELANI I., Computer vision applications in construction: current state, opportunities & challenges, Automation in Construction, 132, (2021)
  • [3] SPENCER B F, HOSKERE V, NARAZAKI Y., Advances in computer vision-based civil infrastructure inspection and monitoring [ J], Engineering, 5, 2, pp. 199-222, (2019)
  • [4] ZAKERI H, MOGHADAS N F, FAHIMOFAR A., Image based techniques for crack detection, classification and quantification in asphalt pavement: a review, Archives of Computational Methods in Engineering, 24, 4, pp. 935-977, (2017)
  • [5] CHA Y J, CHOI W, BUYUKOZTURK O., Deep learning-based crack damage detection using convolutional neural networks [ J ], Computer-Aided Civil Infrastructure Engineering, 32, 5, pp. 361-378, (2017)
  • [6] ATHA D J, JAHANSHAHI M R., Evaluation of deep learning approaches based on convolutional neural networks for corrosion detection, Structural Health Monitoring, 17, 5, pp. 1110-1128, (2018)
  • [7] YANG Na, WANG Dejia, LI Tianhao, Et al., Deformation monitoring of ancient buildings based on computer vision method, Journal of Building Structures, 44, 1, pp. 192-202, (2023)
  • [8] DAWOOD T, ZHU Z, ZAYED T., Machine vision-based model for spalling detection and quantification in subway networks, Automation in Construction, 81, pp. 149-160, (2017)
  • [9] HAN Xiaojian, ZHAO Zhicheng, Structural surface crack detection method based on computer vision technology, Journal of Building Structures, 39, pp. 418-427, (2018)
  • [10] CHA Y J, CHOI W, SUH G, Et al., Autonomous structural visual inspection using region-based deep learning for detecting multiple damage types, Computer-Aided Civil Infrastructure Engineering, 33, 9, pp. 731-747, (2018)
12