摘要
目前,彩色图像饱和度与明度的关系没有明确的定量描述.文中基于Weber-Fechner定律,推导了HSV颜色空间明度与饱和度的定量关系.在自然对数梯度场对大量图片的明度、饱和度运用假设检验方法进行验证,以出现概率最大的值作为斜率的初始假设,计算直线族的截距;将饱和度自然对数值代入直线方程得到假设的明度值,计算与真实明度值之间的误差进行检验;根据误差大小改变假设值,重复上述过程直至满足误差要求.文中得到的明度自然对数值与饱和度自然对数值之间的线性关系模型不具有设备依赖性,对CCD相机和CMOS相机均适用.该模型将彩色图像的像素向量从3维降到2维,降低了数据分析的复杂度,也将为图像传感器的优化设计起到指导作用.
The quantitative relationship between saturation and brightness value of color images had not been determined yet.According to Weber-Fechner law,the relationship equation is proposed.In the gradient field of logarithm value,the saturation and brightness value of a group of color images is analyzed by means of hypothesis-testing.The hypothesis value of slope is set based on the probability distribution.Then the intercepts and the linesequations are got.The termination of the algorithm depends on the error between bright value and the value obtained by the saturation and the equations.The results show that the saturation and brightnesslogarithm values subject to a group of parallel lines with independence from devices.The dimension of pixel vectors of color images could be decreased to two.The proposed model of saturation and brightness value will play very active and important effect of understanding and analysis of color images.And it will have a great influence on the optimal design of image sensors.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2014年第8期1272-1278,共7页
Journal of Computer-Aided Design & Computer Graphics
