3D Reconstruction from Section Plane Views Based on Self-Adaptive Neural Network

In order to represent 3D spatial entity effectively in geological engineering, layered model for geological mass is put forward based on drill hole information. Firstly, for the given geological drill hole data, adaptive neural network is adopted to forecast ore grade of information unknown areas within the geological sections and then geological layered data is obtained. Secondly, based on discretization meshwork model, topological relations for control points can be established automatically between adjacent data layers, so as to construct surface model of 3D spatial entity, which can be visualized by OpenGL technique. Finally, to evaluate the performance of the approach proposed, a 3D simulation system was developed. The experimental results demonstrate that the new modeling method provides a solution to the 3D reconstruction problems existing in the fields without spatial data and can generate complex 3D solid model with higher accuracy and better time performance.