Deep learning–based nondestructive evaluation of reinforcement bars using ground‐penetrating radar and electromagnetic induction data

This paper proposes a nondestructive evaluation method based on deep learning using combined ground-penetrating radar (GPR) and electromagnetic induction (EMI) data for autonomic accurate estimation of the cover thickness diameter reinforcement bars. A real-time object detection algorithm—You Only Look Once–version 3 (YOLO v3)—is adopted to automatically identify bar reflected signals from radargrams, with which range is roughly predicted. Subsequently, EMI data, accompanied range, are imported one-dimensional convolutional neural network (1D CNN), pretrained by calibrated GPR simultaneously estimate diameter. Testing on-site shows that YOLO v3 superior Single Shot Multibox Detector in hyperbolic signal identification. 1D CNN collected an in-house sand pit experiment accuracy is, respectively, 96.8% 90.3% permissible error 1 mm. Further, concrete specimens demonstrates among 22 estimated values (including thickness), there 17 accurately estimated, while inaccurately have up 2 The experimental results show proposed can autonomically evaluate high accuracy.