基于Graph Cuts的木材扫描电镜图像特征提取方法  

Feature Extraction Method of Wood SEM Micrographs Based on Graph Cuts

作　　者：朱佳[1] 汪杭军[2] 

机构地区：[1]浙江农林大学信息工程学院,临安311300 [2]浙江农林大学天目学院,诸暨311800

出　　处：《林业科学》2014年第4期108-114,共7页Scientia Silvae Sinicae

基　　金：国家自然科学基金项目(60970082);浙江省自然科学基金项目(Y13C160027);浙江农林大学人才启动项目(2013FR059)

摘　　要：提出一种基于Graph Cuts图像分割的木材SEM图像特征提取方法。首先对木材SEM图像增强,应用Graph Cuts分割图像;提取木材组织轮廓后,使用数学形态学平滑对象轮廓;最后提出木材细胞的9种几何和统计特征及其计算方法。针对12个木材树种的试验结果表明:本方法给出的特征在给定树种及各树种的早、晚材间均具有较大的差异,为木材材种识别提供了重要依据。Generally,even the subtle differences of wood textures among different species can be clearly identified with their scanning electron microscope（ SEM） micrographs. Therefore,the features extracted from the SEM micrographs can provide a reliable basis for the classification of wood micrographs. In this paper,a novel method of wood SEM micrographs was proposed for feature extraction based on Graph Cuts. Firstly,the wood SEM micrographs would be enhanced before partitioning the images into multiple segments by Graph Cuts,then extracting the contours of wood tissues and smoothing the contours of objects by mathematical morphology. Finally,nine kinds of geometric and statistical features and their computing methods were given. The experimental results on 12 wood species showed that the features we proposed had a great difference among the given wood species（ also in the earlywood and latewood of each given species）,which provided an important basis for the wood assortment recognition.

关 键 词：GRAPH CUTS 特征提取 扫描电镜 图像分割 细胞 

分 类 号：S781[农业科学—木材科学与技术]