Real-time reconstruction of moving objects in an electrical capacitance tomography system using inter-frame correlation

Samir Teniou, Mahmoud Meribout, Khaled Belarbi

Research output: Contribution to journalArticlepeer-review

10 Scopus citations


In this paper, two new electrical capacitance tomography (ECT) algorithms for real-time reconstruction of a sequence of images of moving objects passing through a given two-dimensional section of electrical electrodes are presented. They both explore the inter-frame correlation that usually exists between consecutive frames and use this information as an additional constraint during the reconstruction of the actual frame in solving both its forward and inverse problems. In addition, contrary to the Kalman filter-based tomography technique, both suggested methods maintain the intrinsic non-linearity property of the inverse and forward problems. The proposed algorithms were tested using both synthetic and real ECT data for some moving objects and by adding up to 7% of noise in the values of capacitances. The corresponding experimental results demonstrate that the use of inter-frame correlation for ECT image reconstruction can lead to a significant improvement in image quality, execution time, and stability. Hence, a processing time of up to 175 frames/s was achieved on a single processor PC compatible computer. This interesting performance would be even higher if a dedicated parallel hardware architecture implements some intrinsically parallelizable, though time consuming, subroutines required by the proposed methods, such as matrix multiplication and matrix inversion, is used.

Original languageBritish English
Article number6184268
Pages (from-to)2517-2525
Number of pages9
JournalIEEE Sensors Journal
Issue number7
StatePublished - 2012


  • Basis constraint method
  • electrical capacitance tomography
  • finite element method
  • motion estimation
  • moving objects


Dive into the research topics of 'Real-time reconstruction of moving objects in an electrical capacitance tomography system using inter-frame correlation'. Together they form a unique fingerprint.

Cite this