AN ERT IMAGING ALGORITHM BASED ON NELDER- MEAD SIMPLEX METHOD

Electrical resistance tomography (ERT) techniques have increasing gained importance during the past several decades as a process visualization tool. Most of the existing ERT methods work in term of an optimization process and assume the optimized objective function be differentiable. But this assumption doesn't hold in noisy conditions and leads to an infeasible optimization process. In this paper, a new ERT imaging method based on NeIder-mead simplex algorithm is proposed. The key feature of the algorithm can efficiently work without the differentiability constraint. This strategy can efficiently enhance spatial resolution with less calculation than the existing algorithms. A group of experiments show that the proposed method can work better than two existing ET imaging methods in complex conditions with noisy data.