A TRAINABLE GRADING SYSTEM FOR TOBACCO-LEAVES

A grading system based on image processing techniques was developed fo r automatic inspection and grading of flue-cured tobacco leaves. The s ystem used machine vision in extraction and analysis of color, size, s hape, surface texture and vein. A two-dimensional feature space was pr oposed to express feature distribution of tobacco leaves. The space wa s found to be well confined in an elliptic region. A database was cons tructed to record the feature distribution of standard contrast tobacc o leaves prepared by experts through visual evaluation. The decision o n grades was made based on the so-called 'nearest-neighbor' method for which the overall difference among features between the measured toba cco leaves and the standard contrast samples were used as a target par ameter for judgment. This system can be easily trained by users with t he knowledge of the feature distribution information of different toba cco leaves. It can also be adapted for the inspection of other agricul tural products.