Research Of The On-line Detection Of Stem Ratio In The Tobacco Leaf Based On X-ray Transmitission Imaging

In the process of threshing and redrying,the percentage of stem in leaves and the percentage of large and medium flakes are key quality control indicators.According to the relevant regulations,the percentage of stem in leaves after threshing and removing is less than 2.0%.However,in the process of threshing and redrying,there is a certain contradiction between the percentage of stem in leaves and the percentage of large and medium pieces,because the percentage of stem in leaves increases with the increase of the percentage of large and medium pieces.Therefore,it is necessary to take into account both the rate of large and medium lobes and the rate of stem content in leaves.At present,the traditional method is still used to detect the stem content in leaves.The traditional method is complex,time-consuming and destructive to the structure of tobacco leaves.It can not verify the repeatability and stability.It is very important for retorting enterprises to explore new methods,new ways and real-time detection of stem content in leaves.Therefore,aim at these problems,the on-line detection technology of stem content in Leaves Based on X-ray transmission imaging will be developed.Based on the transmission principle of X-ray and the attenuation law of X-ray after passing through the substance,the quality of tobacco stem in tobacco slices is fitted,and the stem content in leaves is calculated.This study will provide ideas and preliminary research for exploring and developing new methods to detect stem content in leaves.This paper mainly carries out the following research:(1)The basic principle of X-ray is studied.From the generation of X-rays and the analysis of the properties of X-rays,the principle of X-ray transmission imaging and the attenuation law of X-rays are studied.Through the study of the basic principles,its results show that the idea of this paper is feasible.(2)In the process of image preprocessing,image segmentation technology is mainly studied.In this paper,image-based tobacco leaf spatial model is used to segment tobacco leaf image.It is found that the mean clustering segmentation under color space model is better for tobacco leaf image segmentation.In the specific application,the ratio judgment method of color space is used to segment the single color object.(3)The characteristics of tobacco stem is analyzed.According to the different material properties and attenuation properties of tobacco stem,the difference of gray level on X-ray images is shown.The diameter of tobacco stem is calculated by skeleton method and bisected connected region method,and classified according to the diameter.The development of related software is completed to realize the automatic classification and identification of the stems.(4)After determining the experimental conditions,off-line and on-line tests are carried out.The experiments are carried out mainly using pure stem,free leaf mixed with free stem and cigarette smoke of different quality,and repeatability and accuracy experiments are done respectively.The repeatability mainly verifies the stability,and the accuracy experiment mainly verifies the feasibility of the method.After a large number of tests,the repeatability coefficient of pure stem experiment is less than 5%,the repeatability coefficient of tobacco online experiment is less than 8%,the maximum accuracy error of offline experiment is less than 8%,and the maximum accuracy error of online experiment is less than 10%.The experimental results are satisfactory and meets the detection requirements.