To solve the problem of the "soft-field" effect and the proximity of sensitivity matrix in electrical capacitance tomography technology, an image reconstruction algorithm based on least squares support vector machine (LS-SVM) and particle swarm optimization (PSO) has been developed, namely LSP algorithm. Group image samples of some typical flow patterns are trained, and the capacitance deviation generated by the proximity of the sensitivity matrix is predicted. On the basis of the deviation,

more »

... fitness function in particle swarm optimization is established. The nonlinear adjustment method for inertia factor and the mutation operation of particles speed are used in the standard particle swarm optimization. The improved PSO algorithm is used to search for the optimal solution of the reconstructed image. The simulation shows that compared with Landweber algorithm, the image precision of LSP algorithm is improved nearly 50%, and the image error decreases to less than 20% for some typical flow patterns. Therefore, LSP algorithm can be applied to the non-real-time ECT system with high precision. This paper presents a feasible and effective way to research on image reconstruction algorithm for ECT.