计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (28): 158-161.

• 图形、图像、模式识别 • 上一篇    下一篇

根系CT序列图像区域生长分割的新方法

陈郁淦,周学成,乐 凯   

  1. 华南农业大学 工程学院 南方农业机械与装备关键技术省部共建教育部重点实验室,广州 510640

  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-10-01 发布日期:2011-10-01

New method for region growing segmentation of plant root CT image sequences

CHEN Yugan,ZHOU Xuecheng,LE Kai   

  1. Key Lab of Key Technology on Agricultural Machine and Equipment,Ministry of Education,College of Engineering,South China Agricultural University,Guangzhou 510640,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-10-01 Published:2011-10-01

摘要: 在对传统区域生长算法改进的基础上,针对原位根系CT序列图像的特点,提出了一种基于区域生长的植物根系CT序列图像分割算法。通过对208幅JPEG格式的植物根系CT序列图像进行直方图分析,确定植物根系区域的分割阈值范围,结合阈值分割实现改进区域生长法对单层根系图像进行分割得到目标区域。在此基础上,利用植物根系在介质环境中的空间连续性,进一步实现仅在选择单幅图像种子点的情况下一次性完成整套CT序列图像的分割。借助MITK(Medical Imaging Toolkit)工具箱对分割好的原位根系CT序列图像进行三维重建,对三维模型进行不同角度观测来判断分割的正确性。实验结果表明,该算法分割速度快、精度高,能够有效地去除CT图像背景中杂质像素,准确提取出植物根系目标区域。

关键词: 原位根系, CT图像, 空间连续性, 区域生长, 序列图像分割

Abstract: A new algorithm based on the improved region growing segmentation method is put forward to segment plant root CT image sequences in this paper.The 208 plant root CT image sequences are analyzed with histogram analysis,and the basic threshold of segmentation for plant root region is defined.With threshold segmentation,the improved region growing method is implemented to extract the root region from CT image.After that,according to the continuity of root system in the space of soil,the whole sequences of CT images are segmented for once under the condition of choosing seed points on single image only.With the toolkit of MITK(Medical Imaging Toolkit),3D reconstruction of plant root CT image sequences segmented is implemented to check the correctness of segmentation by observing the 3D model of plant root from different views.Experimental results show that this algorithm is faster of running speed and higher of segmenting accuracy,with which the impurity pixel existing in the background of CT image can be removed effectively and the root region is extracted correctively.

Key words: plant root in situ, CT image, spatial continuity, region growing algorithm, sequence images’segmentation