The paper presents further progress of using magnetic induction tomography (MIT) for image reconstruction of the target objects under examination. The forward problem is a typical eddy current problem in electromagnetic computation. This problem can be formulated through a diffusion equation in terms of the magnetic vector potential and solved using the edge finite element (FE) method. The induced voltages in sensing coils are then computed by taking the line integral of the magnetic vector potential around the coil loop. It is evident that the induced voltages are a non-linear function of the electrical conductivity for non-magnetic conductive materials. Sensitivity analysis can be achieved, which reflects the changes of the induced voltage with respect to variation in conductivity. The inverse problem is to reconstruct the conductivity distribution using the measurement data, which is an ill posed and non-linear problem. Non-linear reconstruction methods are usually used to obtain the desirable images. The regularization technique can be used to assist in the stable solution through incorporating as much prior knowledge as possible. The MIT image maps the electrical conductivity distribution, hence the shape of the flow. For the image reconstruction we applied an inverse finite element technique that is able to reconstruct fully 3D MIT images of the electrically conductive objects. The experimental tests have been conducted and the images reconstructed with both experimental and simulation data presented.