Inverse problem: comparison between linear back-projection algorithm and filtered back-projection algorithm in soft-field tomography

An image reconstruction algorithm in process tomography is very important in ensure that the reconstructed image is satisfied. The back-projection algorithm is the most popular algorithm applied in process tomography including a linear back-projection (LBP) algorithm and a filtered back-projection (FBP) algorithm. The objectives of this paper are: (1) to compare the result from LBP and FBP algorithm in soft-field tomography, and (2) to investigate the effect of FBP algorithm to a reconstructed image in soft-field tomography. Also, a comparison between hard-field tomography (ultrasonic tomography) and soft-field tomography (electrical resistance tomography, ERT) had been discussed in this paper. The ERT had been conducted based on a simulation using COMSOL Multiphysics live link with MATLAB software. As a result, a nature behavior of softfield gave a main factor of inaccurate position of reconstructed image using FBP algorithm. When it multiplied with the concentration profile from LBP result, a “hill” surface sensitivity distribution and a color scale shifted in FBP gave inaccurate result. Therefore, it is believed that it gave a reason of why the FBP algorithm is not the concern on reconstructing image in soft-field tomography