Annotated Impact Echo Dataset (IE2020)

Using artificial intelligence, deep learning in particular, requires datasets that are accurately annotated. An annotated impact echo dataset (IE2020) is formed to benchmark, develop, and apply deep learning models for concrete bridge deck evaluation using non-destructive evaluation and testing methods. IE2020 includes impact echo waveforms stored in text files where the first column represents the time and the second column represents the amplitude of the signal. IE2020 is segmented in two folders, each contains impact echo waveforms collected from defected (D) and sound (S) concrete. IE2020 is the first dataset of its kind that is validated with the ground truth (known location and geometry of the defects).