CAJ | 학술논문

针对现有电阻抗成像算法的局限，在此将修正的粒子群算法与牛顿拉夫孙结合算法，形成MPSO-MNR算法。对二维圆形求解区域，采用有限元剖分，在三角电流驱动模式下，应用提出的MPSO-MNR算法进行电阻抗重构，并研究了噪声对重构结果的影响。数值仿真结果表明：MPSO-MNR算法能够准确重构解域内电阻抗分布；噪声影响成像的质量，随着噪声的增加（信噪比的减少），重构目标的边界、背景变得逐渐模糊。MPSO-MNR算法避免要求迭代初值接近真值，并具有较快收敛的特点，在一定的噪声范围内，可用于电阻抗重构。
침대현유전조항성상산법적국한，재차장수정적입자군산법여우돈랍부손결합산법，형성MPSO-MNR산법。대이유원형구해구역，채용유한원부분，재삼각전류구동모식하，응용제출적MPSO-MNR산법진행전조항중구，병연구료조성대중구결과적영향。수치방진결과표명：MPSO-MNR산법능구준학중구해역내전조항분포；조성영향성상적질량，수착조성적증가（신조비적감소），중구목표적변계、배경변득축점모호。MPSO-MNR산법피면요구질대초치접근진치，병구유교쾌수렴적특점，재일정적조성범위내，가용우전조항중구。
Because of the limitations of the existing algorithms of the electrical impedance tomography(EIT),the MPSO-MNR algorithm is formed by combining the modified particle swarm optimization and the modified Newton-Raphson algorithm. The electrical impedance distribution is reconstructed by the MPSO-MNR algorithm,and the influence of the noise on the recon-struction results is analyzed,where the 2-D circular solution domain is subdivided by finite element method(FEM),the current is injected by trigonometric current driven pattern. The numerical simulation results show that the MPSO-MNR algorithm can re-flect the electrical impedance distribution in the solution domain,the noise influences the imaging quality,and the border and the background of the reconstruction objects become vague gradually with the increase of the noise. The MPSO-MNR algorithm avoids the demand of the iterative initial value close to the true value and has the faster convergence speed. It can be used to the EIT within the range of certain noise.