Assessment of Deep Learning Methods for Concrete Crack Detection and Classification

  • P. Jayesh Kumar, A. Diana Andrushia, N. Anand

Abstract

Cracks are the common indication factor for the structural safety analysis.  In Structural Health Monitoring (SHM) concrete crack detection is the important research field. Therefore, it is mandate to study the automation methods of crack detection. Many image processing methods and machine learning methods are involved in this field of research. The deep learning methods which are in structural damage analysis are still in infancy stage. This paper presents the detail review of deep learning methods for concrete crack detection and classification. It deals with different deep learning networks, deep learning software’s and architectures for crack detection and classification in the recent years. It also summarizes the performance metrics of various state-of-the-art deep learning methods. At the end, the clear direction for deep learning-based automation techniques are presented. Deep learning-based methods will be helpful for the automation engineers to build the powerful automotive system to check the health of the structures.

Published
2020-03-30
How to Cite
P. Jayesh Kumar, A. Diana Andrushia, N. Anand. (2020). Assessment of Deep Learning Methods for Concrete Crack Detection and Classification . International Journal of Advanced Science and Technology, 29(3), 13342 -. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/31534
Section
Articles