应用兰德韦伯算法的木材缺陷图像重建  

An Reconstructed Image Algorithm for Wood Defect Based on the Landweber Method

作　　者：刘嘉新[1] 高景泉 李超[1] Liu Jiaxin;Gao Jingquan;Li Chao(Northeast Forestry University,Harbin 150040,P.R.China)

机构地区：[1]东北林业大学

出　　处：《东北林业大学学报》2019年第12期125-128,共4页Journal of Northeast Forestry University

基　　金：国家自然科学基金项目(31700643)

摘　　要：以含有缺陷的红松木材和含有空洞的蒙古栎木材为样本,应用兰德韦伯(Landeweber)迭代算法、代数重建法(ART)对原木缺陷图像重建,比较2种方法的差异,遴选木材无损检测相对较好的方法。结果表明:在迭代50次情况下,红松木材的缺陷图像,Landweber算法相对误差率23%、图像拟合度75%,ART算法相对误差率83%、图像拟合度53%;蒙古栎木材的缺陷图像,Landweber算法相对误差率10%、图像拟合度91%,ART算法相对误差率17%、图像拟合度79%;与ART算法相比,Landweber算法重建缺陷的图像能够精确反映缺陷位置,图像拟合度较高,证明Landweber算法适用于木材应力波缺陷图像重建。Taking Pinus koraiensis wood with defect and the hollow Quercus mongolica wood as the experimental material,we applied the Landweber iterative algorithm and ART iterative algorithm to reconstruct the wood defect images.The better method is determined for nondestructive testing of wood by comparing those two techniques.The experimental results show that,for the defect image of P.koraiensis wood,the relative error rates of Landweber algorithm and ART algorithm are 23%and 83%,respectively,while the degrees of image fitting are 75%and 53%,respectively.As for the defect image of Q.mongolica wood,the relative error rates of Landweber algorithm and ART algorithm are 10%and 17%,respectively,while the degrees of image fitting are 91%and 79%,respectively.It is observed that the position of defect can be located more accurately from the reconstructed image with the Landweber algorithm than that with the ART algorithm,and the degree of image fitting of Landweber algorithm is higher than that of ART algorithm,demonstrating that Landweber algorithm is suitable for image reconstruction of wood stress wave defects.

关 键 词：木材无损检测 木材缺陷图像 应力波检测 兰德韦伯迭代算法 代数重建法 

分 类 号：S781.5[农业科学—木材科学与技术]