烟叶褐变表观颜色值及其判别函数研究

Study on color space data–based discriminating functions for tobacco leaf browning

以100份陈化的湖北烤烟烟叶样品正反面5种颜色值(亮度值、红度值、黄度值、饱和度值、色调角)为指标,采用逐步判别分析的方法,对不同褐变档次烟叶样品进行判别分析,建立判别函数。结果表明:烟叶正面亮度值L、红度值a、饱和度值c和色调角h进入判别函数并建立判别模型。用自身验证法和交互检验法对原样品进行回判,回判准确率为99.0%,对25个验证样品的判别准确率达到100%。

To classify the browning degree of tobacco leaves, the stepwise discriminant analysis was used to discriminate the different browning grades of tobacco leaves on the positive and negative sides of 100 aged Hubei tobacco leaves based on 5 color values including brightness(L), red(a), yellowness(b), saturation value(c) and hue angle(c), and the discriminant function was established. The result showed that color value L, a, c and h on the positive of tobacco leaves entered the discriminant function. Self–verification and cross–examination were used to judge the original samples, the total accuracy was 99.0%, and the discriminant accuracy for the 25 validation samples reached 100%.