Detection of Concrete Structural Defects Using Impact Echo Based on Deep Networks
Juncai Xu, Xiong Yu
Abstract
Abstract Deep learning is widely used in image processing, which significantly improves the performance of image classification detection. Based on the current status of concrete structure defect detection technology, this experimental study on the detection of concrete structure defects using impact echo was conducted. Focusing on the unsteady features of the impact echo signal, we adopted wavelet transforms at different scales to extract the wavelet spectrum. At the same time, the convolution and subsample operation were combined to establish the recognition system of concrete structure defect detection based on the deep learning network. The research results show that this system can accurately recognize defects in the concrete structure and has high detection accuracy in the concrete structure assessment process.