Spectral parameters hybrid conjugate gradient method in elec-tromagnetic tomography technology
Electromagnetic tomography is a kind of process detection technology based on the principle of electromagnetic induction.Image reconstruction is the key to solve its inverse problem.Based on the conjugate gradient algorithm,the search direction and iteration parameters are modified.A hybrid conjugate gradient algorithm based on spectral parameters is derived to improve the quality and convergence of image reconstruction.The laboratory model and sensitivity distribution matrix of the electromagnetic tomography technology system are introduced as the basis of the inverse problem.A new spectral parameter conjugate gradient algorithm is proposed to modify the search direction,which is used to control the angle between the old and new search directions.The search direction is determined according to the residual of each step in order to find the optimal solution.Combining the advantages of FR algorithm and PRP algorithm in nonlinear conjugate gradient algorithm,they are mixed by certain proportion to obtain a new hybrid conjugate gradient algorithm.The spectral parameter hybrid conjugate gradient method is applied to the electromagnetic tomography technology laboratory system,three typical laboratory models are constructed,and the spectral parameter hybrid conjugate gradient algorithm and the traditional algorithm are compared and evaluated.The experimental results show that the spectral parameter hybrid conjugate gradient algorithm has higher quality of reconstructed image and better numerical performance,and it combines the advantages of FR algorithm and PRP algorithm,with faster convergence speed and higher imaging quality than other algorithms.
NETL
NSTL
万方数据
