摘要
通过图像扫描仪获得的彩色图像常包含有一些样稿颜色以外的杂色或噪声颜色，当用基于一维直方图的传统的颜色统计法进行颜色分割时，会出现错分现象。对此，本文提出了基于二维直方图的颜色分割新方法，实现了一些ＣＡＤ系统图像颜色的准确分割和归并。实验结果和实际应用都证明了新方法的有效性，该方法也可应用于全真彩色图像的量化。
Color images which are obtained from image scanners contain many colors beyond the original colors. When the conventional statictical algorithms based on one dimensional histogram are used for color segmentation, some colors can be misclassified. To remedy this problem, a new method for color segmentation based on two dimensional histogram is prosented, which can be used to accurately segment and merge colors for color images in some CAD systems. The experimental results and some real applications show the effectiveness of the proposed method. The method is also used for color quantization of real color images.
出处
《光学技术》
CAS
CSCD
1999年第4期50-51,共2页
Optical Technique
关键词
颜色分割
二维直方图
颜色聚类
CAD
color segmentation
two dimensional histogram
color clustering
