摘要
在分析木材材色分布特征的基础上,提出针对木材图像HSV颜色空间的色调、饱和度和亮度3分量分别进行非等间隔量化降维,根据木材图像色调分布主要在0°～60°区域内,饱和度主要在0～0.6,亮度主要在0.66～1的区间上,将色调量化为9个非等间隔区间,饱和度量化为3个非等间隔区间,亮度量化为3个非等间隔区间。并在此基础上采用主色调的方法,将3个颜色分量合成一维特征矢量,从而进行相似性比对检索。结果表明,该方法能够保证检索准确率,同时提高检索效率。
A non equal interval quantization method was proposed in the light of hue, saturation and value of wood image in HSV color space based on the analysis of wood color distribution feature. The hue is quantized into 9 non equal intervals, the saturation into 3 non equal intervals, the value into 3 non equal intervals according to the hue distribution of wood image mainly in 0°- 60° , saturation in 0-0.6, and brightness in 0.66-1. And based on this, the three color components are combined into one dimensional feature vector for similarity retrieval using the primary color method. Experimental results show that this method can ensure the retrieval accuracy, at the same time, improve the efficiency of retrieval.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
2012年第10期159-162,共4页
Journal of Northeast Forestry University
基金
国家林业局"948"项目(2011-4-04)资助
关键词
木材图像
非等间隔
主色调
Wood images
Non equal intervals
Dominant color
