Self-adaptive marching cubes 3D surface reconstruction method for 3D-GIS spatial raster volume data

Abstract

According to the spatial information, to establish the 3D surface model for underground geological structure is an important research point in the 3D-GIS field. The 3D-GIS spatial raster volume data are serial slices of 2D raster data along the Z-axis, and the data set can generate the geological model by 3D surface reconstruction method. The traditional marching cubes (MC) algorithm only considers the isolated cubes of the volume data, and it may produce “ladder” 3D surface topology that cannot meet the geological requirements. In this paper, a self-adaptive marching cubes 3D surface reconstruction method is proposed. The method adopts the 3D region growing technique to process the spatial raster volume data, and set the Region of Interest (ROI) weights to each raster point, and uses a specific spatial template to calculate the ROI density of the cube. Based on the calculation, the MC cube can self-adaptively segment into the MC sub-cube, which can build fine 3D surface according to the surrounding influence. The experimental results show that the improved method can reconstruct smoother 3D surface model than traditional MC algorithm.