Vision-based concrete crack detection using a hybrid framework considering noise effect

Article Engineering, Civil

Experimental and numerical investigation on the role of interface for crack-width control of hybrid SHCC concrete beams

Shozab Mustafa et al.

Summary: Through experimental and numerical studies, it was found that stronger interfaces enable composite action in hybrid beams and provide better crack width control compared to artificially weakened interfaces.

ENGINEERING STRUCTURES (2022)

添加到收藏夹

Article Engineering, Multidisciplinary

Crack detection of concrete structures using deep convolutional neural networks optimized by enhanced chicken swarm algorithm

Yang Yu et al.

Summary: This study proposes a vision-based crack diagnosis method using deep convolutional neural network (DCNN) and enhanced chicken swarm algorithm (ECSA) for model training and optimization. The method is tested on image patches cropped from damaged concrete samples, and its performance is evaluated using a group of statistical evaluation indicators.

STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL (2022)

添加到收藏夹

Article Construction & Building Technology

Data-driven damage assessment of reinforced concrete shear walls using visual features of damage

Sina Mansourdehghan et al.

Summary: This paper proposes a damage assessment framework based on the visual features of a damaged reinforced concrete shear wall, and validates it using experimental data. Different machine learning techniques are used for classification, and predictive equations are proposed to estimate the performance level, residual strength, and drift ratio of the walls. The results show that the Random Forest model is the most efficient method, and the predictive equations can accurately estimate the peak drift ratio and residual strength.

JOURNAL OF BUILDING ENGINEERING (2022)

添加到收藏夹

Article Construction & Building Technology

Cost-effective system for detection and quantification of concrete surface cracks by combination of convolutional neural network and image processing techniques

Pengyong Miao et al.

Summary: This study proposed a semi-automated system for crack detection and quantification, using a trained CNN and an application. GoogLeNet was determined as the CNN model for the study, with transfer learning showing balanced performances on two datasets. The study introduced a new sliding window technique and crack width calculation method, and developed an application integrating these methods to detect and analyze cracks effectively.

CONSTRUCTION AND BUILDING MATERIALS (2021)

添加到收藏夹

Article Engineering, Civil