光照强度对烟叶颜色特征向量的影响

烟叶颜色特征是用于烟叶等级判别的重要因子,影响烟叶图像质量和等级判别的重要因素是光照强度(简称光强)。为研究光强如何影响不同等级的烟叶颜色特征,设计烟叶图像采集系统,以人工分级所用光源光强为基准,在12个不同光强下分别采集不同等级烟叶图像,提取RGB、HIS及HSV模型的颜色分量,分析各等级的烟叶颜色分量在不同光强下的变化规律,研究不同等级各颜色分量对光强的敏感性,提出适合于烟叶等级判别的颜色分量和光强,为烟叶自动分级及光源选取提供科学依据。试验结果表明,对于R、V颜色分量,光强为2500 lx左右时,判别不同烟叶等级效果最佳。G、B、I和HSV模型中HHSV分量,光强为3200 lx左右时,判别不同烟叶等级效果最佳。

Influence of light intensity on color features of the flue-cured tobacco leaf

Color feature is an important factor affecting the automatic tobacco leaf grading system. Light intensity affects tobacco leaf image quality and grade criteria. To study how light intensity influences tobacco color features, we designed an image acquisition system in which pictures were taken under twelve different light intensities based on the artificial light intensity. Using color vectors under the model RGB, HSI and HSV to analyze how light intensities influence color vectors and sensibility. Suitable color vectors and light intensity for grading tobacco leaves were identified, providing a scientific basis of the light selection for tobacco classification. The result showed that color vectors R and V can be used as discriminant operators under the light intensity of 2500 lx, while color vectors G, B, I and H of the model HSV can be used as discriminant operators under the light intensity of 3200 lx.