摘要
针对木材表面活节缺陷对象,提出一种基于被囊群体算法(Tunicate Swarm Algorithm,TSA)与模板匹配的死节缺陷图像阈值分割算法。首先,输入RGB彩图转换成三通道(R通道、G通道、B通道)灰度图像,分别对各通道灰度图进行处理。然后将正态分布、泊松分布、瑞利分布的模型匹配度作为目标适应度函数,将像素范围的搜索空间中的取值定为候选解,候选解通过进化迭代计算得到全局最优解。最后通过阈值分割得到二值图。结果表明,算法中G通道泊松分布性能最优,Ja、Di、Bf指数平均值分别为75.60%、86.35%、78.95%。
A threshold segmentation algorithm based on the matching of tunicate swarm algorithm(TSA)and template matching is proposed for wood surface defects.Firstly,RGB color image is input to transform into three channels(R channel,G channel and B channel)gray image,and each channel gray image is processed respectively.Then,the model matching degree of normal distribution,Poisson distribution and Rayleigh distribution are taken as the target fitness function,and the pixel value in the search space within the pixel value range is taken as the candidate solution,which is obtained by evolutionary iteration.Finally,binary image is obtained by threshold segmentation.The results show that the G-channel Poisson distribution has the best performance,and the average values of JA,Di and BF are 75.60%,86.35%and 78.95%,respectively.
作者
李健
褚超
郭康乐
黄元
程玉柱
LI Jian;CHU Chao;GUO Kang-Yue;HUANG Yuan;CHENG Yu-Zhu(College of Mechanical and Electronic Engineering,Nanjing Forestry University,Nanjing 210037,China)
出处
《林业和草原机械》
2020年第3期48-52,共5页
Forestry and Grassland Machinery
基金
南京林业大学大学生创新项目(2018NFUSPITP161
2019NFUSPITP0171)
关键词
木材缺陷
被囊群体算法
模型匹配
阈值分割
wood defect
tunicate swarm algorithm
template matching
thresholding segmentation
