Efficient Two-Dimensional Blocked Element Compensation

Very large, two-dimensional, anisotropic arrays have been proposed to improve ultrasound image quality. Due to noncontiguous acoustic windows into the body, however, a significant portion of such an aperture may be blocked. Blocked elements result in high sidelobes in the point spread function, degrading image quality. To compensate for this, an object dependent method using multiple receive beams has been recently proposed. This method is effective in removing undesired sidelobes. However, previous results were for one-dimensional arrays where only lateral beams were used for estimation. With two-dimensional arrays, the distribution of blocked elements can change beam characteristics, both laterally and elevationally. In other words, receive beams must be formed in both directions for better performance. Although straightforward in principle, extension of the algorithm from one dimension to two increases computational complexity dramatically. Furthermore, the restricted elevational steering capability of anisotropic arrays also limits performance. In this paper, several computationally efficient algorithms for two-dimensional blocked element compensation are proposed and evaluated. It is shown that undesired sidelobes can be effectively removed using only a limited number of receive beams. Image quality can therefore be restored in the presence of blocked elements without significantly increasing hardware complexity.