计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (2): 211-215.DOI: 10.3778/j.issn.1002-8331.1810-0147

• 图形图像处理 • 上一篇    下一篇

基于颜色量化和密度峰聚类的彩色图像分割

王鹏宇,游有鹏,杨雪峰   

  1. 南京航空航天大学 机电学院,南京 210001
  • 出版日期:2020-01-15 发布日期:2020-01-14

Color Image Segmentation Based on Color Quantization and Density Peak Clustering

WANG Pengyu, YOU Youpeng, YANG Xuefeng   

  1. College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210001, China
  • Online:2020-01-15 Published:2020-01-14

摘要: 彩色图像分割是簇绒地毯数字化制造的关键技术,图像的分割质量直接影响到后续的图像处理。为解决地毯的彩色图像分割问题,针对人眼在RGB颜色空间中感知不均匀的特性,提出了一种基于颜色量化和密度峰聚类的彩色图像分割算法。基于Lab颜色空间进行颜色量化,在HVC颜色空间中用NBS距离来衡量人眼对颜色差异的感知程度,采用改进的密度峰聚类算法自动确定聚类中心,从而分割地毯图案。实验结果表明,该算法能在不影响人眼感知的前提下得到颜色种类少且边缘清晰的地毯分割图像。

关键词: Lab颜色空间, NBS距离, 密度峰聚类, 图像分割

Abstract: Color image segmentation is the key technology of tufted carpet digital manufacturing. The quality of image segmentation directly affects the subsequent image processing. In order to solve the problem of carpet color image segmentation, a color image segmentation algorithm based on density peak clustering and color quantization is proposed to overcome the inhomogeneous perception of human eyes in RGB color space. Firstly, color quantization is performed based on Lab color space, and then NBS distance is used to measure the difference of color perception in HVC color space. Finally, density peak clustering algorithm is used to automatically determine the clustering center and segment the carpet image. Experimental results show that the algorithm can get carpet segmentation images with few color categories and clear edges without affecting human perception.

Key words: Lab color space, NBS distance, density peak clustering, image segmentation