沥青混合料图像矢量化建模的改进方法

Improving method of image vectorization modeling for asphalt mix ture

针对沥青混合料数字图像矢量化建模问题,提出采用拆分合并强度方法对沥青混合料轮廓曲线进行多边形近似,并以均方根误差作为迭代计算收敛判据.采用自编程序完成了多边形迭代逼近计算.与现有的矢量化方法相比,本方法避免了多边形近似中对尺寸的依赖效应,对于构建多尺度的沥青混合料模型具有更好的稳定性,有效地减小了分析模型的规模.

To explore the DIP-FEM（digital image processing-init e element method） modeling procedure of asphalt concrete,split-and-merge stre ngth method was presented to reconstruct the outline based on polygonal approxim ation.Root mean square error（RMSE） was used as the iterative convergence crite rion.Self programming was used to calculate polygon interation approximation.C o mpared with existing methods,the proposed method can eliminate the size-depend ent effects of polygonal approximation to construct a multi-size model more sta bly,and the scale of analyzing the model can be effectively reduced.