基于CT图像的土壤孔隙结构重构
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:

北京市科技计划项目(Z161100000916012)、中央高校基本科研业务费专项资金项目(2015ZCQ-GX-04)、国家自然科学基金项目(41501283)、国家重点研发计划项目(2017YFD0600901)和北京市共建项目


Reconstruction of Soil Pore Structure Based on CT Images
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    土壤孔隙的几何结构和空间特征决定了土壤的透气性和保水性,对土壤功能多样性和生态修复具有重要影响,但现有对土壤孔隙的研究中,缺乏可直观性和定量性对孔隙特征进行描述的工具和方法。针对这一问题,本文采用基于面绘制的移动立方体法(Marching cubes, MC)和基于体绘制的光线投射法(Ray casting, RC)还原土壤孔隙的几何形态和空间分布。以单个孔隙和不同孔隙密集程度的土壤孔隙CT图像为应用对象进行实验,结果表明,2种算法的重构效果均不受土壤样本孔隙密集程度的影响。其中,MC算法重构出的孔隙结构存在边界锯齿化和缺失的现象,且其孔隙体积也小于实际情况;而RC算法重构的孔隙轮廓清晰,结构真实,可完整地呈现出孔隙结构的细节信息。为进一步评价2种算法的重构性能,采用模型品质、绘制速度和内存消耗3个指标进行实验结果的比较分析。结果表明,MC算法存在二义性的不足,使得孔隙结构存在一定程度的失真,重构的孔隙模型质量一般,但由于其只针对表面体素进行重构,因而具有较快的绘制速度和较小的内存消耗;而RC算法采用为每个体素分配不透明度和光强的方法来合成模型,避免了MC算法的缺点,能够保持孔隙模型的细节信息,但由于其重构过程中所有体素点都参与运算,使得其绘制速度较慢,内存占用较大。通过对模型品质、绘制速度和内存消耗3个指标的综合分析,RC算法更加适用于土壤孔隙的三维重构,不仅为土壤孔隙的可视化分析提供了一种较为先进的方法,也为研究土壤水分和养分的运移以及空气的交换奠定了技术基础。

    Abstract:

    The geometric and spatial characteristics of soil pores determine the permeability and water retention of soils, which have important effects on soil functional diversity and ecological restoration. However, in the existing research on soil pores, there is a lack of tools and methods for describing intuitive and quantitative characteristics of pores. To solve this problem, the three-dimensional reconstruction technique based on marching cubes algorithm and ray casting algorithm was used to restore the geometry and spatial distribution of soil pores. The objects of experiment were the CT images of soil pores with a single large pore and different pore densities. The comparison results of the two algorithms showed that the pore structures of different sizes reconstructed by MC algorithm had boundary aliasing and missing phenomena, and the pore volume was also smaller than the actual situation. While the RC algorithm reconstructed the pore contour clearly and the structure was real, the detailed information of each pore structure can be completely represented, and the reconstruction effect was not affected by the porous density of the soil sample. In addition, in order to further evaluate the reconstruction performance of the two methods, three indicators of model quality, rendering speed and algorithm characteristics were used to compare and analyze the experimental results. In order to ensure the comparability of the refactoring indicators, all experiments were carried out by using the VTK graphics development library, based on the Visual Studio 2017 programming platform of the same computer. The results showed that the MC algorithm was only for the reconstruction of the pore surface voxels, so it occupied less memory and had faster drawing speed. However, due to the ambiguity of the algorithm, the quality of the reconstructed pore model was general. The RC algorithm calculated all the voxel points of the sample, so it took up a large memory and drew slowly. However, assigning opacity and light intensity were used to each voxel to synthesize the model, avoiding the disadvantages of the MC algorithm, and it was able to maintain the details of the pore model. In summary, the RC algorithm would be able to provide a more advanced method for the visualization of soil pores, laying a technical foundation for studying soil moisture and nutrient transport and air exchange.

    参考文献
    相似文献
    引证文献
引用本文

赵玥,刘雷,韩巧玲,赵燕东.基于CT图像的土壤孔隙结构重构[J].农业机械学报,2018,49(s1):401-406. ZHAO Yue, LIU Lei, HAN Qiaoling, ZHAO Yandong. Reconstruction of Soil Pore Structure Based on CT Images[J]. Transactions of the Chinese Society for Agricultural Machinery,2018,49(s1):401-406.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2018-07-15
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2018-11-10
  • 出版日期: