| 摘 要: 针对磁感应断层成像(MIT)因病态性和不适定性导致的重建图像伪影多、分辨率低的问题,提出一种基于共轭梯度算法与维纳滤波相结合的 CG-Wiener融合图像重建算法。该算法结合共轭梯度法优化迭代方向以加速收敛,并利用维纳滤波在频域中抑制噪声干扰、增强信号特征。为验证算法性能,基于电磁场仿真软件构建多种电导率分布模型,并与线性反投影(LBP)算法、Tikhonov正则化算法以及传统共轭梯度算法的重建效果对比。实验结果表明:所提融合算法在相同噪声条件下显著提高成像质量,它的重建图像相关系数较传统共轭梯度算法最高提升9.5%,图像误差最高降低4.1%。该算法有效提升了 MIT重建的质量。 |
| 关键词: 磁感应断层成像 灵敏度矩阵 共轭梯度 维纳滤波 图像重建 |
|
中图分类号: TP23
文献标识码: A
|
| 基金项目: :国家自然科学基金项目资助(32071904);浙江省自然科学基金项目资助(LY20C130008) |
|
| Research on Magnetic Induction Tomography Image Reconstruction Based on Conjugate Gradient and Wiener Filter Fusion Algorithm |
|
ZOU Tao, FU Xiaping
|
(School of Information Science and Engineering, Zhejiang Sc-i Tech University, Hangzhou 310018, China)
107366806@qq.com; fuxp@zstu.edu.cn
|
| Abstract: To address the problems of numerous artifacts and low resolution in reconstructed images of Magnetic Induction Tomography (MIT) caused by il-l posedness and il-l conditioning, a CG-Wiener fusion image reconstruction algorithm based on conjugate gradient and Wiener filtering is proposed. This algorithm combines the conjugate gradient method to optimize the iterative direction for accelerated convergence and utilizes Wiener filtering in the frequency domain to suppress noise interference and enhance signal features. To verify the algorithm’s performance, various conductivity distribution models were constructed based on electromagnetic field simulation software, and the reconstruction results were compared with those of the Linear Back Projection (LBP )algorithm, Tikhonov regularization algorithm, and the traditional conjugate gradient algorithm. Experimental results show that the proposed fusion algorithm significantly improves imaging quality under identical noise conditions. Compared to the traditional conjugate gradient algorithm, the correlation coefficient of the reconstructed image is increased by up to 9.5% , and the image error is reduced by up to 4.1% . This algorithm effectively enhances the quality of MIT image reconstruction. |
| Keywords: magnetic induction tomography sensitivity matrix conjugate gradient Wiener filtering image reconstruction |
