Deep Learning-Based Corroded Crack Identification in Reinforced Concrete Using SCNet Model

Abstract

In order to improve the efficiency and accuracy of corroded cracks detection and classification in reinforced concrete, a corroded cracks identification model Steel Corrosion Net (SCNet), based on deep learning Convolutional Neural Network (CNN), is proposed. The SCNet combines massive initial data with a multi hidden layer neural network framework, and achieves feature learning and accurate classification through model training. The data set of 39000 crack figures is firstly built by original data collection and data enhancement. The training process of the SCNet consists of defining the loss function, the selecting back propagation optimization algorithm, continuously entering data to the network framework and running back the propagation algorithm until the error drops to a certain range. Afterwards, a SCNet three-classification neural network model is built and tested using TensorFlow learning framework and Python. According to the training and testing accuracies of the model, the structure and parameters of the SCNet network are optimized. The results of SCNet are compared with those obtained by two traditional testing methods. The results show that the proposed SCNet model achieves a classification accuracy of 96.8%. In other words, it can effectively identify and classify the corroded cracks in reinforced concrete, with high accuracy and measurability.