多重分形理论在原木图像处理中的应用  

Application of Multifractal Theory to Wood Image Processing

作　　者：韩书霞[1] 戚大伟[1] 

机构地区：[1]东北林业大学,哈尔滨150040

出　　处：《东北林业大学学报》2011年第10期132-133,共2页Journal of Northeast Forestry University

基　　金：黑龙江省自然科学基金重点项目(ZD201016)

摘　　要：用计算机断层扫描技术对木材进行无损检测,采用多重分形频谱分析方法对原木CT图像进行边缘检测。首先计算图像各像素点的Hlder指数α(x,y),然后估计出其多重分形频谱f(α)的值,对图像的像素点进行分类;根据多重分形理论,找出图像的边缘点。结果表明:多重分形频谱分析方法对木材CT图像的边缘检测具有良好的效果,并具有较好的局部性。The computer tomography （CT） technology was applied to wood nondestructive testing. A method for edge detection of wood CT images was given by multifractal analysis. The Holder exponent of image pixels was computed first, then its multifraetal spectrum was estimated and different image pixels were classified. The image edge points were set according to the multifractal theory. Experiment result shows that this edge detection method for wood CT images has good detection performance and spatial locality.

关 键 词：多重分形 Hlder指数 边缘检测 木材CT图像 

分 类 号：S781.5[农业科学—木材科学与技术]