In non-destructive testing, standard methods reconstruct 3D images by post-processing data from ultrasound probes. However, calculation times - ranging from a few minutes to an hour, depending on the number of voxels in the 3D image - make it impossible to obtain an instant image. A thesis carried out at List, a CEA Tech institute, has overcome this drawback by reducing the calculation time required to obtain the image by a factor of 30.
Starting from existing technologies in the field of medical imaging, researchers at CEA List, a Carnot institute, have developed an algorithm based on spatio-temporal Fourier transforms to take account of the specificities of non-destructive testing: diversity of waves, presence of interfaces between the probe and the part to be tested, reflection of waves by interfaces... An initial proof of concept verified the validity of the algorithm by parallelizing calculations on multi-core graphics processors. According to CEA List, the 3D images obtained are of equivalent quality to those obtained by conventional methods, with calculation times reduced to a few seconds.
The algorithm enables the geometry of defects to be visualized directly in three dimensions with a single sensor position. A new thesis aims to accelerate data transfer times to the computing units for a prototype real-time 3D imager.
Towards more automated and interconnected 3D metrology
Metrologic Group reminds us that metrology 4.0 refers to the integration of digital technologies, automation and Industry 4.0 concepts into measurement and control processes...