TY - JOUR
T1 - Cracks in bridge floor detected by 2-dimensional complex discrete wavelet packet transform
AU - Zhang, Zhong
AU - Hamata, Ohzora
AU - Akiduki, Takuma
AU - Mashimo, Tomoaki
AU - Saito, Taiki
AU - Hayashi, Kazuhiro
N1 - Funding Information:
Acknowledgment. This work was partially supported by JSPS KAKENHI Grant Numbers 16H03143 and 19H02223. The authors also gratefully acknowledge the helpful comments and suggestions of the reviewers.
Publisher Copyright:
© 2020, ICIC International. All rights reserved.
PY - 2020
Y1 - 2020
N2 - Bridge inspection is usually done visually on-site, and for this reason is problematic in places inaccessible to inspectors. In this paper we propose a solution through the analysis of features in captured images that suggest potential problems. We focus on the ability to detect cracks, and a complex-valued Haar wavelet is designed and applied to a 2D-CWPT, which is then combined with an anisotropic diffusion filter. A detection method for linear cracks is proposed through interpolation of the crack loss area and extraction processing according to the shape feature. The main results are: 1) we confirm that the complex-valued Haar wavelet is more effective for edge enhancement than the complex-valued Meyer wavelet; 2) quantitative evaluation of the proposed method for crack detection using the correct rate, sensitivity, specificity, precision rate, and F-value confirms the effectiveness of the method.
AB - Bridge inspection is usually done visually on-site, and for this reason is problematic in places inaccessible to inspectors. In this paper we propose a solution through the analysis of features in captured images that suggest potential problems. We focus on the ability to detect cracks, and a complex-valued Haar wavelet is designed and applied to a 2D-CWPT, which is then combined with an anisotropic diffusion filter. A detection method for linear cracks is proposed through interpolation of the crack loss area and extraction processing according to the shape feature. The main results are: 1) we confirm that the complex-valued Haar wavelet is more effective for edge enhancement than the complex-valued Meyer wavelet; 2) quantitative evaluation of the proposed method for crack detection using the correct rate, sensitivity, specificity, precision rate, and F-value confirms the effectiveness of the method.
KW - 2D-CWPT
KW - Complex-valued Haar wavelet
KW - Crack
KW - Feature extraction
KW - Fre-quency analysis
KW - Image processing
UR - http://www.scopus.com/inward/record.url?scp=85096031297&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85096031297&partnerID=8YFLogxK
U2 - 10.24507/ijicic.16.06.2007
DO - 10.24507/ijicic.16.06.2007
M3 - Article
AN - SCOPUS:85096031297
SN - 1349-4198
VL - 16
SP - 2007
EP - 2019
JO - International Journal of Innovative Computing, Information and Control
JF - International Journal of Innovative Computing, Information and Control
IS - 6
ER -