Convolutional Neural Network Based Defect Detection Method for Concrete Structures
Influenced by the environment and loads,structural defects often occur in concrete structures,affecting the stability and integrity of the structure.The traditional method of detecting structural defects in concrete structures is through manual detection,which is an inefficient and labor-intensive process.In order to realize the intelligent detection of concrete structural defects,a detection method based on YOLOv7 deep learning algorithm is proposed.Firstly,a concrete structure defects dataset is established,the number of dataset is expanded by image processing,and the concrete structure defects are labeled.YOLOv7 algorithm is chosen to detect the concrete structure defects,and the optimization and improvement of the algorithm is carried out.Finally,a test platform based on deep learning is constructed,and the defects are detected in the defective images by using the completed algorithm after training.The test results show that the model can basically detect all the concrete structural defects,and has a relatively good ability to recognize the concrete defects under the conditions of rain,covered with dust and so on.
NETL
NSTL
万方数据
