Improved PSO ERT image reconstruction algorithm for human lung based on prior knowledge and clustering

Abstract

Aiming to improve the resolution of the reconstructed image in electrical resistance tomography (ERT) technique for human lung, an improved particle swarm optimization (PSO) image reconstruction algorithm is proposed based on prior knowledge and clustering, according to the characteristics of ERT technique for human lung. In the initial iteration of the novel algorithm, the modified Newton-Raphson algorithm is adopted to solve the inverse problem of ERT for human lung. According to the image reconstruction result of the modified Newton-Raphson algorithm, the finite elements in the sensitivity field are clustered according to the resistivity, based on the prior knowledge. On that basis, the improved PSO image reconstruction algorithm is adopted to the reconstruction of the image. Simulation results demonstrate that, compared to the modified Newton-Raphson algorithm and the PSO image reconstruction algorithm based on prior knowledge, the proposed algorithm can improve the precision of image reconstruction effectively.