Research On Adaptive Meshing Strategy Based On Rock Surface Texture Structure

The modeling of rock surface texture structure is of great significance for characterizing rock texture characteristics,studying rock sample destruction rules,and simplifying image complexity.Currently,mesh modeling cannot achieve a good balance between the number of meshes and the accuracy of the model.Most of the accuracy of the model is improved by continuously increasing the number of meshes,while controlling the number of meshes to improve post-processing efficiency often loses the accuracy of the model,and models with high accuracy often adopt the method of increasing the mesh density,which cannot form the rock texture boundary.Therefore,it is of great theoretical and practical significance to propose a modeling method that can balance the number and accuracy of meshes and has a good expression of rock texture boundaries.Voronoi diagram,also known as Voronoi meshing,is a very important branch in computational geometry.The final results of the Voronoi diagram are all convex polygon meshes.This feature is very helpful for the post-processing calculation of the mesh,and its polygon mesh structure has certain advantages in rock structure characterization.Therefore,this paper proposes to use the Voronoi diagram to build the model of rock texture images.The image is fused with the Voronoi grid.The Voronoi grid is adaptively adjusted according to the image characteristics.Finally,a Voronoi grid segmentation model with image basis and controllable complexity is obtained.The method in this paper makes full use of a variety of Voronoi diagram generation methods,image processing techniques and gradient descent methods.First,the original rock texture image is processed into a feature edge image that is beneficial to post-processing through image processing techniques,including image graying,threshold segmentation,and Canny edge detection.Then calculate the distance of the edge image to generate the corresponding image density map.The density map converts the original pixel-level image into a digital image for initial seed point spreading and initial Voronoi grid segmentation model generation.Then,using the gradient descent method,the Voronoi grid is fused to the rock texture image.According to the characteristics of the pixel changes in the gray map of the rock texture,the direction of the fastest pixel change is found in each iteration,and the appropriate change size is selected.The Voronoi grid model is used for boundary fitting,so that the final meshing model can well express the rock texture boundary.Finally,the proposed Voronoi grid quality evaluation system is used to evaluate the quality of the segmentation results from multiple dimensions,which proves that the method proposed in this paper has a high feasibility in the segmentation of the texture image of the rock surface.The method proposed in this paper can well link the pixel-level image with the geometrically expressed Voronoi grid,so that the final model has strong image basis and can restore the original rock image well.The final result also shows that the method in this paper has a strong adaptive adjustment ability,which can complete the matching of the Voronoi grid and the edge of the image feature according to the image feature,and there is no particularly high requirement for the number of seed points,which is well implemented The balance between model accuracy and model complexity is provided to provide guidance for model postprocessing.At the same time,another important content proposed in this article-the Voronoi grid quality evaluation system quantitatively grades the grid division results from two levels and four dimensions.