In order to improve the ability of ultrasonic guided wave technology in accurately detecting damage location, size and shape in plate-like structures, the low-resolution imaging results of the total focusing method(TFM) and the multi-scale deep learning algorithm model are combined to study a high-precision imaging method for targeting two type of damage:round holes and cracks.The high-precision imaging algorithm based on deep learning is composed of two parts: a convolutional neural network and a deconvolutional neural network. The multi-scale analysis, nonlinear enhancement and multi-level fusion functions of the neural network are used to improve the resolution. On the basis of the imaging results of the all-focus imaging algorithm, a network training database is constructed, and the trained network is used to detect and verify the two cases of circular hole-crack double damage and circular hole-circular hole double damage in the plate. The results show that the imaging method has high imaging accuracy, and can further obtain the fine features of the damage based on the precise location of the damage.

ultrasonic guided wave; plate-like structures; total focusing method(TFM); deep learning