基于SFS的土壤三维结构恢复及孔隙度计算  

3D Reconstruction and Porosity Computing of Soil Based on SFS Algorithm

作　　者：王虓[1] 温维亮[1] 郭新宇[1] 王传宇[1] 

机构地区：[1]北京农业信息技术研究中心,北京100097

出　　处：《中国农业科学》2013年第23期4950-4956,共7页Scientia Agricultura Sinica

基　　金：国家自然科学基金项目(31171454);国家863计划课题(2013AA102404);国家科技支撑计划课题(2012BAD35B01)

摘　　要：【目的】为了快速、准确地计算土体结构和孔隙度等特征,提出一种基于SFS算法(shape from shading)的土壤三维结构重构方法。【方法】首先获取土壤的灰度显微图像,根据朗伯表面漫反射模型建立土壤的光照方程;然后利用泰勒展开法和雅克比迭代法求解光照方程,计算出像素点高度值并实现土壤的三维重建;最后根据正态分布校正高度值计算土壤孔隙度。【结果】实验结果证明,该方法计算得出的孔隙度与环刀烘干法测得的结果相差0.81%,误差率1.75%。【结论】利用显微图像中土壤结构的三维信息,可以客观准确地计算出土壤孔隙度等物理特征。[ Objective ] In order to calculate the characteristics of soil structure and porosity quickly and accurately, a method based on SFS algorithm （shape from shading） is proposed for 3D reconstruction of soil. [Method] This method deduced the illumination equation based on Lambertian surface reflection model, and then solved this equation for working out the pixels＇ height of soil＇s gray microscope images with the method of Taylor expansion and Jacobi iteration. [Result] Experimental results demonstrate that compared with the soil porosity measured by soil sampler and baking, this method has a result with 0.81% less and 1.75% error rate. [Conclusion] The proposed method, by using soil microscope images＇ 3D information, could calculate the soil physical characteristics like porosity more objectively and accurately.

关 键 词：土壤微形态学 SFS 三维结构重建 土壤孔隙度 正态分布 

分 类 号：S152[农业科学—土壤学] TP391.41[农业科学—农业基础科学]