Nondestructive Detection of Stem Content in Tobacco Strips Using X-Ray Imaging Analysis

For tobacco strips used in cigarette production, the stem content is an important quality index to assess the impurity level of the cut leaves. The presented work developed a nondestructive detection method of stems in cut leaf agricultural products by the low energy X-ray imaging. The algorithm of stem image processing and weight calculation principle was established, and then a machine vision system with X-ray imaging and image analysis was set up to verify the quantitative detection method. The results showed that the relative error of the detection method ranged from −3.64% to 2.76%. The determination of stems with a different morphology, such as the thick stem, were also realized based on the image analysis. The accuracy of determining thick stem and long stem was 94.67% and 99.33%, respectively. The developed method is superior to the current ISO detection method of tobacco stem in leaves under the same testing conditions in terms of accuracy and efficiency, which could be applied as an effective online detection method to monitor the quality of processed leaf for cigarette production.