Research And Realization Of The Online Detection Of Stem Ratio In The Tobacco Leaf Based On X-Ray Image

Now in tobacco redrying industries,the measurement of the stern ratio in tobacco leaf mainly use traditional method.The first step is manual sampling,then realize the separation of the stem and the leaf by threshing and air separation method.Next step is screen the stem by vibration and screening equipment,after this step,weigh the separated stem in order to calculate the stem ratio in the tobacco leaf.Aim at this low efficiency,poor accuracy and destructive detection method,this thesis proposes a method that use machine vision technology to detect the stem ratio in tobacco leaf by non-destructive and online method,and put into practical use in industry,effectively increase the detection accuracy,meanwhile economic and reliable.This thesis mainly refer to research and realization of the online detection of stem ratio in the tobacco leaf based on X-Ray image.The major work of this thesis includes:Firstly,we researched the possibility that use hyperspectral scheme to detect the stem ratio in the tobacco leaf,and we took some experiments,information collection and information analysis,and deduced that this scheme is not suitable for the detection of the stem ratio.Then we researched the possibility that detect the stem ratio in tobacco leaf based on X-Ray image,analyzed the feasibility and advantage of this scheme.Secondly,we researched the process that use X-Ray image to realize the detection of stem ratio in tobacco leaf.We studied and realized the process of the image processing,includes the image smoothing,image segmentation,low-pass filter,connected components labeling,skeleton extraction and thinning,confirm the diameter of the stem and calculate the area of the stem with different diameter.Thirdly,we developed a complete system to detect the stem ratio in tobacco leaf online and real-timely,we also developed a set of software to detect to stem in tobacco leaf based on C++,includes image processing program,engineer control module,real-time show and so on,and at last we put it into practical use in redrying industry.Fourthly,we installed and debugged the site plant,completed camera calibration,and optimized the equipment.After that,we conduct some experiments,includes comparison test,repetitive test and so on,and proved that this system can be in normal operation in practical use,and satisfied the industry's requirement.Compared to traditional detection method,it can significantly improve the efficiency and accuracy.