Estimation of High Velocities in Synthetic-Aperture Imaging—Part II: Experimental Investigation

The paper describes the performance of a new pulse sequence design and estimation approach for increasing the maximum detectable velocity in synthetic aperture (SA) velocity imaging. Measurements are conducted for conventional imaging for comparing the velocity range detectable by a directional Transverse Oscillation (TO) autocorrelation estimator to a new cross-correlation estimator. For conventional focused emissions a 192-elements, 3 MHz convex array transducer is used together with the SARUS experimental scanner on a flow rig at beam-to-flow angles of 60°, 70° and 90°. Here the new estimator always yields a higher precision, and the aliasing limit is increased by a factor 3. The new SA inter-spaced scheme was investigated using Field II simulations and SARUS measurements. A 3 MHz, 128-elements phased array was employed with a 5 virtual source emissions scheme for flow estimation and 15 emissions for B-mode imaging. The scheme was interleaved three times for a positive, negative, and positive transmission, so that non-linear pulse inversion also could be made. The experiments were conducted at three angles and for 4 different pulse repetition frequencies. A peak transverse velocity of 0.51 m/s could be estimated at fprf=450 Hz, translating to 5.6 m/s at fprf=5 kHz showing the theoretical increase of a factor 10 predicted in the accompanying theory paper.