计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (2): 183-188.

• 图形图像处理 • 上一篇    下一篇

基于阈值及像素聚类的分割算法性能比较

杨雪丹,刘文萍   

  1. 北京林业大学 信息学院,北京 100083
  • 出版日期:2014-01-15 发布日期:2014-01-26

Comparison of algorithms and performance of thresholding and clustering segmentation

YANG Xuedan, LIU Wenping   

  1. College of Information, Beijing Forestry University, Beijing 100083, China
  • Online:2014-01-15 Published:2014-01-26

摘要: 图像分割是图像处理到分析的关键步骤,阈值分割方法因其计算简单而被广泛应用,聚类算法也因其准确性成为图像分割领域中一类极其重要的算法。选取几种经典阈值分割算法和几种聚类算法对几幅毫米波图像进行分割实验,并引入错分类误差、均匀测度、区域间灰度对比度作为算法测评标准,比较了各种算法对毫米波图像的分割性能。

关键词: 聚类算法, 图像分割, 阈值分割, 测评

Abstract: Image segmentation is the key step to do image processing and image analysis. Thresholding segmentation method is widely used for its simpleness and the clustering algorithm also plays an important role in the image segmentation field. Some millimeter wave images are processed by certain classic thresholding segmentation algorithms and several clustering ones. As evaluation standard, misclassification error, uniformity measure, gray-level contrast are implemented to compare segmentation performance of each algorithm.

Key words: clustering algorithm, image segmentation, thresholding segmentation, evaluation