Abstract:

3D vector field interpolation is an important preprocessing step before topology analysis and visualization. There are two shortcomings of the inverse distance weighted (IDW) method for 3D vector field interpolation. On the one hand, the selection of reference samples lacks of scientific guide. On the other hand, the local behavior of the vector field cannot be expressed. Considering these two shortcomings, an adaptive IDW interpolation method involving local behavior for 3D vector field is proposed. Firstly, the original sample set is searched based on the distance impact degree. Then, the original set is classified as small sample set or large sample set. In the case of the small sample set, it can be used for the IDW method directly. However, in the case of the large sample set, the optimal sample set for the IDW interpolation method can be acquired by the optimization step based on the local linear approximation of the vector field. Theoretical analysis and experiment results indicate that the proposed method can select the reference samples adaptively, and the interpolation result is more reasonable due to the assumption of local linear approximation of the vector field in the local range. The experiment results also show that proposed method improve the interpolation accuracy.