Abstract: The quality of concrete directly impacts the structural safety and service life of buildings. Traditional concrete quality management methods, which rely heavily on manual visual inspection, are inefficient, subjective, and prone to errors, failing to meet the demands of managing quality issues in large-scale engineering projects. Additionally, real-time monitoring and rapid response to concrete quality issues present significant challenges, as traditional management methods often fail to timely detect problems and implement effective remedial actions. This study enhances onsite workers' understanding and repairs efficiency of concrete defects by utilizing the classification results of deep learning as triggers for augmented reality (AR), clearly displaying concrete defect feature models, detailed defect information, and solutions within an AR virtual view. Furthermore, the identified concrete defects are associated with the specific concrete ID, and the defect records are automatically generated and updated, which further improves the automation and information level of quality management, and facilitates the follow-up quality tracking and analysis.