Abstract: Since the sensitivity field in the capacitance sensor of electrical capacitance tomography system is “soft field”, and the “soft field” nature is ignored by the traditional image reconstruction algorithms, there is bottleneck in improving the imaging accuracy for the algorithms. To solve the problem, based on the analysis of the distribution of sensitivity field and the discussion of the “soft field” effect and its impact on the image reconstruction, a novel image reconstruction algorithm is proposed, which is dual particle swarm collaborative optimization. In the algorithm, to eliminate the impact generated by ignoring the “soft field” nature, a priori condition is used to construct the fitness function of particle swarm optimization. The priori conditions under the different flow patterns are obtained by the least square support vector machine. Meanwhile, by introducing the Lotka-Volterra model, a new cooperative-competitive scheme is discussed. The diversity of particles is increased by intraspecific and interspecific learning and competition. So the algorithm improves the global convergence and convergence rate. The experimental results illustrate that this algorithm not only has higher image precision and stronger convergence, but also is resistant to the interference of noise in the measurement signal.

Key words: electrical capacitance tomography, image reconstruction, particle swarm optimization, Lotka-Vloterra model, priori conditions