Efficient imaging techniques using an ultrasonic array

Over the past few years, ultrasonic phased arrays have shown good potential for non-destructive testing (NDT), thanks to high resolution imaging algorithms that allow the characterization of defects in a structure. Many algorithms are based on the full matrix capture, obtained by firing each element of an ultrasonic array independently, while collecting the data with all elements. Because of the finite sound velocity in the specimen, two consecutive firings must be separated by a minimum time interval. Therefore, more elements in the array require longer data acquisition times. Moreover, if the array has N elements, then the full matrix contains N-2 temporal signals to be processed. Because of the limited calculation speed of current computers, a large matrix of data can result in rather long post-processing times. In an industrial context where real-time imaging is desirable, it is crucial to reduce acquisition and/or post-processing times. This paper investigates methods designed to reduce acquisition and post-processing times for the TFM and wavenumber algorithms. To reduce data capture and post-processing, limited transmission cycles are used. Post-processing times is also further reduced by demodulating the data to baseband, which allows reducing the sampling rate of signals. Results are presented so that a compromise can be made between acquisition time, post-processing time and image quality. Possible improvement of images quality, using the effective aperture theory, is discussed. This has been implemented for the TFM but it still has to be developed for the wavenumber algorithm.