Application of crack detection algorithm using convolutional neural network in concrete pavement construction
by Wuqiang Wei; Xiaoyan Xu
International Journal of Grid and Utility Computing (IJGUC), Vol. 13, No. 2/3, 2022

Abstract: Concrete pavement cracks detection is specifically studied based on the application principle of Convolutional Neural Network (CNN) in large-scale image recognition and processing. Firstly, the feature extraction and selection principle are introduced for sample data of concrete pavement crack, and the sample Image Processing (IP) method is expounded. Secondly, the pre-processing of concrete pavement crack images is proposed. Through the establishment of the functional relationship between Crack Rate (CP) and Crack Rate Index (CRI), the samples are trained, and the crack detection model based on CNN is implemented. The CNN model is evaluated by comparing the accuracy and loss rate of the CNN model with the traditional Alex model in processing different numbers of images. Afterward, a concrete pavement crack detection platform is developed with cross-platform Python, OpenCV, and QT framework, combined with Deep Learning (DL), graphical interface development, and image processing. Thereupon, a concrete pavement health evaluation method is proposed. Regression analysis shows that the evaluation method can reasonably evaluate concrete pavement crack.

Online publication date: Tue, 26-Jul-2022

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Grid and Utility Computing (IJGUC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com