Deep Learning-based Crack Detection in the Exterior Walls of Residential Apartments

Dong-Gyun Jung; Jaeho Lee; Yeon-suk Jeong; Junwon Park; Young-Cheol Yoon

doi:10.7734/COSEIK.2025.38.2.75

All Issue

2025 Vol.38, Issue 2 Next Page

Research Paper

Deep Learning-based Crack Detection in the Exterior Walls of Residential Apartments 딥러닝 기반 공동주택 외벽 균열 탐지 정확도 향상에 대한 연구

30 April 2025. pp. 075-084

PDF XML

Abstract

This study compares and analyzes various data preprocessing methods for deep learning-based crack detection in the exterior walls of apartment buildings. It focuses on identifying false-positive crack anomalies not typically represented in standard crack datasets but relevant to apartment buildings. To optimize detection accuracy, two primary image preprocessing techniques were tested: edge filtering combined with color normalization for object detection and image filtering techniques, such as RGB filters. These techniques were applied to differentiate false-positive cracks from genuine wall cracks, and their impact on detection performance was thoroughly evaluated. An EfficientNet V2s-based model was employed to assess performance in detecting false-positive cracks. Four image filters (RGB, YUV, LAB, and HSV) were applied to both original images and those treated by CLAHE normalization. The findings indicated that traditional normalization methods, which are effective for monochrome concrete crack detection, had limited effectiveness in identifying the false-positive crack anomalies of the exteriors of apartment buildings. Additionally, the application of single color filters did not consistently improve detection results. The study concludes that a broader investigation into various combinations of image normalization and color filters was conducted to enhance crack detection performance. Further, a multifaceted approach is necessary to yield better outcomes in detecting real cracks from false positives.

Keywords

deep learning

crack detection

pseudo-crack discrimination

image normalization

EfficientNet V2s model

본 연구는 딥러닝 기술을 활용하여 공동주택 외벽의 균열 탐지를 효과적으로 하기 위한 다양한 데이터 전처리 방법을 비교 분석하였다. 특히, 표준 균열 데이터셋에 일반적으로 나타나지 않는 오탐균열을 식별하는 데 중점을 두고 있다. 이 연구는 탐지 정확도를 최적화하기 위해 여러 이미지 전처리 기법을 적용한 결과를 비교한다. 객체 탐지를 위한 엣지 필터링과 RGB 색 필터 등을 이용한 색상 정규화를 결합한 방법을 집중적으로 검증하였다. 이러한 기술들은 실제 균열과 오탐균열을 구분하기 위해 적용되었으며, 이들의 탐지 성능에 미치는 영향을 철저히 조사하였다. 효율적인 균열 탐지 모델을 찾기 위해 EfficientNet V2s 기반 모델을 적용하였다. RGB, YUV, LAB, HSV 네 가지 이미지 필터가 원본 이미지와 CLAHE 정규화된 이미지에 적용되었는데, 그 결과 단색 콘크리트 균열 탐지에 효과적인 전통적인 정규화 방법이 공동주택 외벽 균열 탐지에는 제한적인 효과를 보인다는 것을 확인하였다. 또한, 단일 색 필터의 적용이 일관된 탐지 결과 개선 효과를 주지 않는다는 것을 밝혔다. 결국, 본 연구를 통해 다양한 이미지 정규화와 색 필터 조합의 균열 탐지 성능을 검증하였으며, 실제 균열과 오탐균열을 구분하는 탐지 성능 향상을 위해 추가적으로 다양한 접근의 연구가 필요하다는 것을 확인하였다.

키워드

딥러닝

균열 탐지

오탐 균열 식별

이미지 정규화

EfficientNet V2s 모델

References

Activeloop (2024) CIFAR-100 Dataset, Retrieved from: https://datasets.activeloop.ai/docs/ml/datasets/cifar-100-dataset/

Cho, H., Yoon, H.J., Jung, J.Y. (2018) Image-based Crack Detection using Crack Width Transform (CWT) Algorithm, IEEE Access, 6, pp.60100~60114.

10.1109/ACCESS.2018.2875889

Choi, I., Shin, K., An, H., Koo, J., Son, J., Lim, D., Oh, T., Yoon, Y. (2020) Development of Image Process for Crack Identification on Porcelain Insulators, J. Korean Inst. Electron. Mast. Eng., 33(4), pp.303~309.

Choi, Y., Kim, S. (2022) Optimization of Image Augmentation Scale Considering Reliability and Computational Efficiency when Classifying Concrete Structure Cracks in CNN, Proc. Korean Inst. Inf. & Commun. Sci. Conf., pp.324~327.

Fairchild, M.D. (2013) Color Appearance Models (3rd ed.), Wiley.

10.1002/9781118653128

Golewski, G.L. (2023) The Phenomenon of Cracking in Cement Concretes and Reinforced Concrete Structures: The Mechanism of Cracks Formation, Causes of Their Initiation, Types & Places of Occur., & Methods of Detect.-A Rev., Build., 13(3), p.765.

10.3390/buildings13030765

Gowda, S., Yuan, C. (2019) ColorNet: Investigating the Importance of Color Spaces for Image Classification, arXiv:1902.00267v1.

10.1007/978-3-030-20870-7_36

Hamishebahar, Y., Guan, H., So, S., Jo, J. (2022) A Comprehensive Review of Deep Learning-based Crack Detection Approaches, Appl. Sci., 12, p.1374.

10.3390/app12031374

Hoang, N.D., Nguyen, Q.L. (2018) Metaheuristic Optimized Edge Detection for Recognition of Concrete Wall Cracks: A Comparative Study on the Performances of Roberts, Prewitt, Canny and Sobel Algorithms, Adv. Civil Eng., 2018(1), p.7163580.

10.1155/2018/7163580

Hutchinson, T., Chen, Z. (2006) Improved Image Analysis for Evaluating Concrete Damage, J. Comput. Civil Eng., 20(3), pp.210~216.

10.1061/(ASCE)0887-3801(2006)20:3(210)

Ji, H., Kim, J., Hwang, S., Kim, D., Park, E., Kim, Y., Ryun, S. (2021) Deep Learning Models for Autonomous Crack Detection System, KIPS Transactions on Software and Data Engineering, 10, pp.161~168.

Keras Modules (2024) Module: tf.keras.application, Retrieved from: https://www.tensorflow.org/api_docs/python/tf/keras/applications

Kim, A., Kim, D., Byun, Y., Lee, S. (2018) Crack Detection of Concrete Structure using Deep Learning and Image Processing Method in Geotechnical Engineering, J. Korean Geotech. Soc., 34(12), pp.145~154.

Kim, B., Kim, G., Jin, S., Cho, S. (2019) A Comparative Study on Performance of Deep Learning Models for Vision-based Concrete Crack Detection according to Model Types, J. Korean Soc. Safety, 34(6), pp.50~57.

Kim, J.-H., Shin, Y.-S., Min, K.-W. (2023). Automatic Crack Detection in Concrete Structures Using Deep Learning and Layered Image Analysis. J. Comput. Struct. Eng. Inst. Korea, 31(3), pp.123~135.

Lakshaymiddha (2024) Crack-Segmentation Dataset, Retrieved from: https://www.kaggle.com/datasets/lakshaymiddha/crack-segmentation-dataset

Lee, S.-Y., Huynh, T.-C., Park, J.-H., Kim, J.-T. (2023). Bolt-Loosening Detection using Vision-Based Deep Learning Algorithm and Image Processing Method. J. Korean Soc. Struct. Eng., 39(2), pp.124~135.

Ministry of Land, Infrastructure and Transport (MOLIT) (2016) Manual of Cracks in Concrete Structure for Maintenance Work such as Defect Repair, Daejun.

Nam, W., Jung, H., Park, K., Kim, C., Kim, G. (2022) Development of Deep Learning-based Damage Detection Prototype for Concrete Bridge Condition Evaluation, J. Civil & Environ. Eng. Res., 42(1), pp.107~116.

Parrany, A., Mirzaei, M. (2021) A New Image Processing Strategy for Surface Crack Identification in Building Structures under Non-uniform Illumination, IET Image Processing, 16, pp.407~415.

10.1049/ipr2.12357

Rouchier, S., Woloszyn, M., Foray, G., Roux, J.J. (2013) Influence of Concrete Fracture on the Rain Infiltration and Thermal Performance of Building Facades, Int. J. Heat & Mass Trans., 61, pp.340~352.

10.1016/j.ijheatmasstransfer.2013.02.013

Shin, J., Won, J. (2023). A Study on the Industry Practitioners' Perceptions for the Activation of AI in the Domestic Construction Sector, J. Korean Soc. Constr. Eng., 45(3), pp.215~227.

Smith, A.R., Lyons, G. (1978). HSL and HSV: Alternative Color Representations, ACM Transactions on Graphics (TOG).

Wikipedia - Color Space (2024) Color Space, Retrieved from: https://en.wikipedia.org/wiki/Color_space

Information

Publisher :Computational Structural Engineering Institute of Korea
Publisher(Ko) :한국전산구조공학회
Journal Title :Journal of the Computational Structural Engineering Institute of Korea
Journal Title(Ko) :한국전산구조공학회 논문집
Volume : 38
No :2
Pages :075-084
Received Date : 2024-10-30
Revised Date : 2024-12-07
Accepted Date : 2024-12-09
DOI :https://doi.org/10.7734/COSEIK.2025.38.2.75

Journal of the Computational Structural Engineering Institute of Korea ISSN:1229-3059(Print) 2287-2302(Online) 한국전산구조공학회 논문집

All Issue