为研究和实现珍珠的自动分选,提出了一种对珍珠的色调及饱和度进行分选的方法.将珍珠图像由RGB颜色模型转换到HIS模型中进行处理,根据,(亮度)分量直方图,通过全局分割提取珍珠光亮区,掩模后得到光亮区H(色调信息)、S(饱和度信息)均值.其中针对色调分选要求,运用了基于RPROP算法的人工神经网络学习方法;针对各色系内色彩的深浅差异,运用了K-均值聚类方法对S均值进行分类.从而实现了先按照珍珠表面色彩进行分类,再将同一色系内的珍珠按照饱和度进行再度分选,最终达到颜色分选要求.

Applying HSL(hue,saturation,luminance)color model transformation technology,the research of pearl luster degree classification has been done,and the pearl image was disposed after transformed it from RGB color model to HSL color model.Then,according to the luminance message that has been received,fuzzy pattern recognition method was adopted to distinguish and classify,and at last the classification of pearl luster degree was realized.Afterwards,by giving correction and instruction,aiming at the evaluation method,the method was further perfected and its applicability was improved.