Performance Comparison of Multiple Convolutional Neural Networks for Concrete Defects Classification

Abstract

Periodical vision-based inspection is a principal form of structural health monitoring (SHM) technique. Over the last decades, vision-based artificial intelligence (AI) has successfully facilitated an effortless inspection system owing to its exceptional ability of accuracy of defects’ pattern recognition. However, most deep learning (DL)-based methods detect one specific type of defect, whereas DL has a high proficiency in multiple object detection. This study developed a dataset of two types of defects, i.e., concrete crack and spalling, and applied various pre-built convolutional neural network (CNN) models, i.e., VGG-19, ResNet-50, InceptionV3, Xception, and MobileNetV2 to classify these concrete defects. The dataset developed for this study has one of the largest collections of original images of concrete crack and spalling and avoided the augmentation process to replicate a more real-world condition, which makes the dataset one of a kind. Moreover, a detailed sensitivity analysis of hyper-parameters (i.e., optimizers, learning rate) was conducted to compare the classification models’ performance and identify the optimal image classification condition for the best-performed CNN model. After analyzing all the models, InceptionV3 outperformed all the other models with an accuracy of 91%, precision of 83%, and recall of 100%. The InceptionV3 model performed best with optimizer stochastic gradient descent (SGD) and a learning rate of 0.001.

Deep Learning for Structural Health Monitoring: Data, Algorithms, Applications, Challenges, and Trends

“…By comparison, the optimal model can be found on the one hand, and on the other hand, whether the newly proposed framework can achieve satisfactory performance can be verified. For example, Arafin et al [132] compared the performance of multiple CNN networks, namely VGG-19, ResNet-50, InceptionV3, Xception, and MobileNetV2. The accuracy, precision, and recall rates of the InceptionV3 model were higher than those of other models.…”

“…Damagesensitive feature extraction can be achieved through image processing techniques, ML, and DL [100], where DL can automatically extract features for image classification, object recognition, and semantic segmentation tasks. DL-based SHM mainly uses structural damage images to classify or segment surface defects to achieve crack detection [130][131][132], bolt loosening detection [47], steel bar exposure detection [49,133], or vehicle recognition through vehicle images captured by road cameras [134,135].…”

Breaking Boundaries in Diagnosis: Non-Invasive Anemia Detection Empowered by AI

“…The advantage of MobileNetV2 is its effectiveness in extracting features utilizing lighter convolution processes, which makes it suited for usage in lowpower devices such as mobile devices [65]. The MobileNetV2 classification approach begins with a feature extraction stage that employs a well-tuned CNN architecture [66].…”

Detecção de Rachaduras em Concreto em Imagens com o uso de Redes Neurais Convolucionais

“…O estudo conduzido por [Arafin et al 2022] focou na avaliação comparativa do desempenho de modelos para a detecção de rachaduras e descamação em superfícies de concreto. Nesse contexto, foram empregadas redes neurais convolucionais (CNNs) pré-treinadas, incluindo VGG-19, ResNet-50, InceptionV3, Xception e MobileNetV2, para a classificação das rachaduras.…”