Real-time 3D imaging with Fourier-domain algorithms and matrix arrays applied to non-destructive testing

Real-time 3D ultrasound imaging with matrix arrays remains a challenge in Non-Destructive Testing (NDT) due to the time-consuming reconstruction algorithms based on delay-and-sum operations. Other algorithms operating in the Fourier domain have lower algorithmic complexities and therefore higher frame rates at the cost of more storage space, which may limit the number of reconstruction points. In this paper, we present an implementation for real-time 3D imaging of the Total Focusing Method (TFM) and the Plane Wave Imaging (PWI), as well as of their Fourier-domain counterparts, referred to as k-TFM and k-PWI. For both types of acquisition, the Fourier-domain algorithms are used to increase frame rates, and they are compared to the time-domain TFM and PWI in terms of image quality, frame rates and memory requirements. In order to greatly reduce their memory requirements, a new implementation of k-TFM and k-PWI is proposed. The four imaging methods are then evaluated by imaging in real time a block of stainless steel containing a 3D network of spherical porosities produced by additive layer manufacturing using a powder bed laser fusion process. •TFM and PWI in the Fourier domain accelerate real-time 3D imaging with matrix arrays.•A low memory cost implementation of Fourier domain. algorithms is presented•A 3D network of porosities made by additive manufacturing is being imaged.•Fourier domain algorithms allow for higher frame rates and lower memory costs.