基于数字图像的水果分形维数特征
DOI:
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:

*湖南省自然科学基金(项目编号:2007JJ6129)、湖南省农业厅重点课题(项目编号:200704A)


Fruits fractal dimension based on digital image
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    为了定量描述仿球形水果表面轮廓间的差异,通过周长—面积法研究了水果分形维数特征。取大小不同的温州蜜柑20个、麻阳冰糖橙17个和红富士苹果18个,设置3种水果花萼面和侧面两个方向上数字图像红色或蓝色分量阈值,将水果区域内外红色、绿色、蓝色分量分别置为0和1,转化成的二值图像克服了Otsu及迭代式自动阈值法边界模糊与水果区域内存在孔眼的特点。以水果区域细化后的边界像素数和区域内像素数分别为水果的周长和面积,借助周长—面积法求得3类水果两个方向的分形维数。试验表明,3类水果2个方向上的分形维数以花萼面分形维数最小,且该方向上水果间的分形维数差异不显著,但侧面分形维数差异明显。这表明,反映仿球形水果的外形特征除花萼面分形维数外,主要指标是侧面分形维数。

    Abstract:

    Quantitative description of the difference of contours of paraspherical fruits is investigated with fractal dimension with perimeter-area method. Test sample sets are 20 Wenzhou mandarin oranges, 17 Mayang bingtang oranges and 18 red fuji apples. Threshold values of red or blue component of fruit calyx and profile image are used to convert the image to binary image. Red, green and blue components in/out fruit region are set to 0/1. This gets over the shortcoming of Otsu and iterative auto threshold method which has blur boundary and eyelet in fruit region. After boundary thinning, pixel numbers of boundary and fruit region are treated as perimeter and area of fruit region. Fractal dimensions of calyx and profile of three kinds of fruits are calculated with perimeter-area method. Test results show fractal dimension of calyx images is smallest among three kinds of fruits in two directions, and deviation of calyx is not significant between different fruits; fractal dimension variates significantly in profile images. This means the shape feature of paraspherical fruit can be described with fractal dimension especially fractal dimension of profile.

    参考文献
    相似文献
    引证文献
引用本文

温芝元.基于数字图像的水果分形维数特征[J].湖南农业大学学报:自然科学版,2010,36(4):478-482.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2010-03-12
  • 最后修改日期:2010-03-12
  • 录用日期:2010-05-31
  • 在线发布日期: 2010-08-13
  • 出版日期:
文章二维码