基于数码相机的印刷品质量检测是未来发展的方向,在颜色检测领域需要首先解决色空间转换的精度问题。采用分子空间的多项式回归法实现了从RGB 颜色空间到CIEL* a* b* 颜色空间的转换,首先把RGB 颜色空间划分成若干个子空间,然后在每个子空间中运用最小二乘法建立多项式模型,对任意RGB 颜色值根据其所在子空间求解对应的多项式方程,即可得到L*a*b*值。实验表明,该方法的转换精度有了很大程度的提高,能够满足数码相机色空间转换的基本要求,为基于数码相机的印刷品质量检测奠定基础。

Printing quality detection by digital camera is the direction of future development, but the problem of color space conversion accuracy need to be solved in the field of color detection. The color space conversion between RGB color space and CIEL*a*b* color space was achieved using polynomial regression based on subspace partition. RGB color space was divided into subspaces and polynomial models were created in subspaces with least square method. L*a*b* values of any RGB values can be obtained by solving polynomial equation according to their own space. The experiment showed that the conversion accuracy increases considerably by using this new method, which can satisfy the requirement of digital camera and lays the foundation for quality inspection of presswork.