Lateral Deconvolution of Ultrasonic Beams

A number of papers have appeared recently describing methods of lateral resolution improvement in ultrasonic images by deconvolution of the amplitude detected cross-beam characteristic. It is the intention of this paper to demonstrate that such methods of resolution improvement are appropriate only for a limited class of targets. In general, the effect of interference between the r.f. signals from multiple targets at differing distances is such that the amplitude beam width obtained for these targets cannot be considered as the linear combination of shifted replicas of the amplitude beam width for an individual target. This is a consequence of the non-linear character of the detection process. In these cases the method of deconvolving the detected responses, although capable of giving reduced spot sizes, makes errors in the position of the images and is susceptible to false images and sidelobes. The phenomenon is demonstrated with computer and experimental data on point targets.