Simulation Study on Influencing Factors of Image Reconstruction Quality Based on UT System

Abstract

As a typical industrial process tomography (IPT) technology, ultrasonic transmission tomography (UTT) is an effective method for cross-section visualization of the gas-liquid two-phase flow pipeline. The quality of the reconstructed image is necessary to evaluate the performance of the method. Sector-diffusion-matrix linear back-projection algorithm (SLBP) algorithm with two filtering algorithms, Hybrid Reconstruction (HR) algorithm and Adaptive Threshold Filtering (ATF) algorithm, were used to reconstruct and post-process the image of the sensing area. A single bubble flow pattern is discussed in this article. Three variables including the thickness of the tube wall, the position of the gas bubble, and the area of the bubble are involved. Image correlation coefficient (ICC), position error (PE), shape distortion (SD), and area resolution (AR) indexes are selected as evaluation criteria of the imaging quality to evaluate the influence of the variables on the imaging quality. ICC was mainly affected by the area of the bubble, and the fluctuation reached 0.5. The relationship between image quality and the three factors was nonlinear. But overall, the thicker the wall, the larger the distance or area, the greater the PE. SD and AR were affected by the bubble area. The larger the area, the smaller the SD. The relationship between image quality and bubble area was also nonlinear. When the bubble area was 314mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> , AR was the smallest, 0.03. When the wall thickness was 3mm, the bubble was located at (-10, 10), and the bubble area was 707mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> , the quality of the reconstructed image was the best, ICC=0.58, PE=0.72, SD=0.96, AR=0.05. The simulation results analyze the influence of the thickness of the pipe wall, the location and area of the reconstruction target on the reconstructed image, and this conclusion has the potential to be applied to industrial processes.