Sparse defects detection and 3D imaging base on electromagnetic tomography and total variation algorithm

Metal products are widely used in the industrial field. However, internal defects such as holes, dents, and scratches are prone to occur due to factors such as processing, production equipment failure, and poor working conditions. Electromagnetic tomography (EMT) is an effective method for defects imaging. Nevertheless, metal defects are prone to be small and sparsely distributed on the surface or inside so that image reconstruction for metal defects based on EMT is still challenging. In this paper, the sparse regularization method is used for a mathematical model of EMT reconstruction in order to improve the image quality. According to the relationship between the detection depth and the excitation frequency, three-dimensional reconstructed images are used for the surface and internal defects of the metal parts. Both simulations and experiments are carried out to verify the effectiveness of the method.