基于模糊模式识别和敏感场优化的ECT图像重建

提出了一种基于模糊模式识别和敏感场优化的电容层析成像(ECT)图像重建方法,以提高重建的精度与质量。首先,提出了一种基于模糊模式的ECT流型识别方法,通过流型识别,选择与输入信号所属相对应的变化敏感场,完成敏感场的优化;然后,提出了一种基于特征提取的敏感场扩充方法,从输入信号中提取特征信息进行数据融合,并通过零填充和随机重组将优化后的敏感场扩展为新的敏感场分布矩阵,进一步优化敏感场;最后,提出一种稀疏图像重建方法,通过构造综合观测方程,求解介电常数分布矢量,并进行图像重建。通过COMSOL软件对ECT系统进行三维仿真建模,仿真实验结果表明,本文方法在图像误差、相关系数等成像指标上优于现有方法,具有更好的成像效果。

Shape Optimization With Sensitivity Analysis in Local Domain

For complicated geometries, it is ineflicient to integrate over the entire domain. A new approach for optimization is presented here, based on sensitivity analysis of local do- main and deformation of full domain. Accurate results of design sensitivity analysis are ob- tained with this approach in shape optimization. This method is shown to be efficient when used in optimization programs and results in less distortion of the mesh.