A Sensitivity Map Deformation Method for Image Quality Improvement of Ultrasonic Tomography

Ultrasonic tomography (UT) technology is a typical industrial process tomography (IPT) technology and an effective method for the cross-sectional visualization of a gas–liquid two-phase flow pipeline. For typical noniterative or iterative algorithms, the design of sensitivity maps is an important aspect for imaging results of UT. The construction of sensitivity maps is often based on the assumption of wave propagation theory; therefore, a sharp edge in a sensitive map (SESM) seems difficult to avoid, which produces a low quality of reconstructed images when the sensing objects appear on the edge. This article proposes the multilevel slope reduction methods to improve image quality via changing the effects of SESM. The multilevel slope reduction method is used to create a gradient slope and to eliminate the SESM effect. The Pearson correlation coefficient (PCC), position error (PE), and area error (AE) are selected as image quality evaluation criteria. The simulation and experimental results show that the proposed method can reduce the AE of the reconstruction results by approximately 0.1. The multilevel slope reduction method can increase the PCCs of the simulation and experimental reconstruction results by 0.02 and 0.06, respectively.