摘要
根据视觉的颜色聚类特性,提出一种图像分割算法.根据从色彩空间(RGB)到(HLC)的变换公式和NBS颜色距离的概念,首先将彩色图像量化为256或更少量化级,而不使图像质量降低;然后依据NBS颜色距离的概念将色彩分类;最后,根据某种规则将颜色区域进行分块,并删除小的颜色块.给出了静物图像和自然景物图像的实验结果.
Proposes an image segmentation algorithm based on perceptual color clustering. Introduces the color space transformation formulae from (RGB) to (HLC) and the NBS color distance. On this basis, color images are quantized to 256 colors or fewer without significantly degrading the color quality and the similar colors are classified based on NBS color distance. Color patterns are finally merged, deleting minor color regions. Experimental results for some color images are given.
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2003年第6期772-775,共4页
Transactions of Beijing Institute of Technology
