Research On Algorithms In Tobacco Sorter System

In the process of cigarette production, tobacco mixes some light and weighty contaminants after baking, re-drying and fermentation. These contaminants directly affect the efficiency of cigarette automated production, and reduce cigarette quality. Currently, the tobacco sorter system based on the spectrum is widely used in the detection of tobacco contaminants. However, there are some problems that are the poor elimination effect of the weighty contaminants and the low tobacco flow in the system. To solve these problems and improve the efficiency of cigarette production, the tobacco sorter system based on multispectral technology and vertical pneumatic separation has been proposed.This paper studies the key technology of the tobacco sorter system based on multispectral technology and vertical pneumatic separation and puts forward the corresponding solutions. The key technology of the system mainly lies in recognition and elimination of tobacco contaminants. In the aspect of tobacco contaminants recognition, an algorithm that is used to solve to the problem of tobacco contaminants recognition is proposed. This algorithm includes three-dimensional modeling based on the moving least square fitting, the storage of the lookup and table online correction of images. In the aspect of tobacco contaminants elimination, an algorithm that is used to improve the elimination rate is proposed. This algorithm controls the elimination of tobacco contaminants by online detecting tobacco velocity and adjusting the elimination delay of the pneumatic electromagnetic valve. It includes image segmentation algorithm, the connected components labeling algorithm and the extraction algorithm of the minimum enclosing rectangle. In the actual test of the system, we test the effect of the two aspects. Experimental results show the solution can achieve our desired effect. The recognition rate of the tobacco contaminants is 97.2%. The elimination rate of the light contaminants is 86.2%.This paper proposes some solutions for this kind of tobacco sorter system and the solutions can effectively improve the performance of the system. The advantages of the solutions mainly lie in the below two aspects. One aspect is building more comprehensive classification model and improving the recognition rate of the contaminants. Another aspect is realizing online adjustment of the elimination delay and improving the elimination accuracy of the contaminants. Therefore, the study has great significance for the application and development of the tobacco sorter system.