计量学报2025，Vol.46Issue(10)：1486-1493,8.DOI:10.3969/j.issn.1000-1158.2025.10.11
基于Calderon引导的深度学习ECT图像重建研究
Research on ECT Image Reconstruction Based on Calderon Guided Deep Learning
摘要
Abstract
To address the distortion and artifacts in reconstructed images caused by the pathology of the inverse problem in capacitance tomography,a multi-scale adaptive residual network structure is designed using a Calderon-guided deep learning method,which can adaptively adjust the size of the convolution kernel to capture the multi-scale information better and reconstruct the images with high accuracy.Experiments demonstrate that the proposed method yields superior results in terms of visualization and objective evaluation metrics compared to sensitivity algorithms,direct imaging algorithms,and convolutional neural networks(CNNs).The experimental results show that the CCORR reaches more than 0.94 using the Calderon pre-imaging results as the input to the multi-scale adaptive residual network,which is more than 4.9%higher than using the capacitance measurements as the input to the network,which reduces the difficulty of extracting the features from the flow pattern,enhances the network recognition ability,and allows for a higher imaging quality.关键词
电学层析成像/流量测量/图像重建/Calderon方法/深度学习Key words
electrical capacitance tomography/flow measurement/image reconstruction/Calderon method/deep learning分类
通用工业技术引用本文复制引用
温丽梅,李成华,山雨泽,马敏..基于Calderon引导的深度学习ECT图像重建研究[J].计量学报,2025,46(10):1486-1493,8.基金项目
国家自然科学基金面上项目(62371452) （62371452）
天津市教委科研计划项目(自然科学)(XJ2023006901) （自然科学）
中国民航大学科研项目(3122024PT06) （3122024PT06）
