Tikhonov正则化
电容层析成像
图像分辨率
迭代重建
平面的
计算机科学
反问题
计算机视觉
断层摄影术
无损检测
人工智能
电容
光学
数学
物理
计算机图形学(图像)
数学分析
电极
量子力学
作者
Ziqiang Cui,Yu Sun,Lifeng Zhang,Huaxiang Wang
标识
DOI:10.1109/tim.2022.3180438
摘要
Planar electrical capacitance tomography (PECT) is sensitive to the dielectric changes in its proximity; therefore, has an attractive prospect in the non-destructive evaluation for the non-metallic composite materials. Currently, the planar ECT employs the static image method for the defect detection, which uses an individual frame of measurements for image reconstruction. The use of static image methods for defect detection depends greatly on the spatial resolution of image reconstruction algorithms. However, the ECT static images are usually of low spatial resolution, primarily due to the ill-posedness in solving its inverse problem. In this paper, a dynamic imaging method has been proposed, aiming to utilizing the information that buried in the consecutive frames. The Tikhonov regularization method is firstly employed for achieving the static image reconstructions. In addition, the level set method has been utilized for the image segmentation to distinguish between the defect and background materials. Subsequently, the dynamic imaging method that based on the frame difference methods has been used for calculating the contour of target defects. The numerical simulations and experiments showed that, the defect of different sizes and shapes could be figured out by using the dynamic imaging method. It also can be shown that, the dynamic imaging method offers more possibility and ways in detecting the defects.
科研通智能强力驱动
Strongly Powered by AbleSci AI