Defect imaging using sub-sampled array data with least squares migration

In ultrasonic phased array imaging, if the array element spacing is greater than half the wavelength, unwanted artefacts known as grating lobes can become prevalent and obscure signals arising from physical targets. This is problematic as use of dense, periodic arrays can result in large quantities of data and an acquisition time that is too lengthy for some applications. Thus, imaging algorithms which can act on sparsely collected data whilst retaining good image quality are highly desirable. Here we apply, for the first time to our knowledge, Least-Squares Migration (LSM), an imaging methodology originating within the seismology community, to sub-sampled ultrasonic array data, resulting in the attenuation of unwanted grating lobes. It is also shown that LSM can be used to obtain improved lateral resolution compared to that achieved by the Total Focusing Method, the current standard in ultrasonic NDT imaging.