木材缺陷的阈值分割算法研究  

Study on Wood Defect Image Threshold Segmentation Algorithms

作　　者：谢永华[1] 陈庆为[1] 

机构地区：[1]东北林业大学机电工程学院,哈尔滨150040

出　　处：《森林工程》2014年第2期38-42,共5页Forest Engineering

基　　金：中央高校基本科研业务专项资金(DL12BB03)

摘　　要：阈值分割算法包括迭代阈值分割、最大方差阈值分割、最大熵阈值分割等,使用3种阈值分割算法进行木材缺陷图像分割,多次实验证明,最大方差阈值分割算法的分割效果最好。在此基础上,对最大方差阈值分割算法进行增强,调整灰度级,增大目标和背景区域的对比度。实验表明,该方法可以明显改善最大方差阈值分割的分割效果。Threshold segmentation algorithms include iterative threshold segmentation, maximum variance threshold segmenta- tion, maximum entropy threshold segmentation and so on. The maximum variance threshold segmentation was the best algorithm among the three threshold segmentation algorithms through repeated experiments. On the basis, by increasing the contrast of the object and the background area, the maximum variance threshold segmentation algorithm was enhanced. The experiments showed that this method could significantly improve the segmentation results of the maximum variance threshold segmentation algorithm.

关 键 词：木材缺陷 阈值分割 迭代阈值分割 最大方差阈值分割 最大熵阈值分割 

分 类 号：S781.51[农业科学—木材科学与技术]