Research On Reconstruction Algorithm And Designment Of Software For ECT System

Electrical capacitance tomography (ECT) is a kind of process tomography (PT) based on capacitance sensor. ECT has the advantage of being non-intrusive, simple in structure, fast in response, low in cost, good insecurity and wide in application. It has gained considerable progress in recent years.There are two typical image reconstruction algorithm in the ECT system at present: linear back-projection (LBP) method and iterative algorithm based on LANDWEBER. LBP algorithm has fast speed in image reconstruction and is fit for real-time image reconstruction, but its reconstruction quality is not accurate. Iterative algorithm can obtain more accurate result through modifying the image repeatedly. But more time consuming is its disadvantage. They have a same feature, that is the positive problem model based on sensitivity matrix, this model assumes any variation of the measured capacitance is caused by proportional variation of dielectric constant on the whole cross section of the pipeline and the sensitivity valuation. The assumption does not conform to reality obviously. The soft market character of capacitive sensor makes the relation between the electrode capacitive and the distribution of the dielectric constant complicated. It can establish the ECT system model through the method of analytical derivation only under some assumption which is unreasonable and unreality. The accuracy and the reasonable of the model can not be promised.This paper does the research on the relation between the electrode capacitance and the distribution of the dielectric constant through multiple linear regression (MLR). That is establishing statistical model of them. It can avoid some necessary assumption which is unreasonable and unreality when establishing model through analytical derivation, then improves the quality of the reconstructed image.This paper establishes the positive problem model of the ECT system through multiple linear regression (MLR), and does the combination with different technologies for solving inverse problem. Two new image reconstruction algorithm are proposed. One of them does combination with general iterative-single step (GIS). That forms a new real-time image reconstruction algorithm (MLRIS) whose quality of image reconstruction is better than LBP algorithm obviously. The other one does the combination with the technologies of LANDWEBER algorithm for solving inverse problem. That forms a new image reconstruction algorithm which has high accuracy of image reconstruction.Increasing the number of projection data can improve image reconstruction quality, however, placing more electrodes around the pipe will produce smaller capacitances and bring detection difficulties to measuring circuit. Image reconstruction for 16-electrode sensor is investigated in this paper.The new algorithm embedded in the software is compiled under the environment of Object-Oriented Visual C++, the software realizes reveiving capacitance values and reconstructing images online, the software has four kinds of images reconstruction algorithm: LBP, iterative algorithm and two kinds of new algorithm proposed in this paper, the ECT system experiment is debugged based on upper and lower computers, then this system runs well and can monitor the changes of medium distribution in the pipe continuously.