The Research Of Tobacco Leaf Automatic Grading System

Tobacco leaf as an important raw material in the tobacco production, its quality level will directly affect the quality level of the tobacco products, so we need to judge and classify the quality grade of tobacco. Tobacco classification are based on artificial subjective judgment in the past, this grading standards are vulnerable to subjective factors, the classification results are not standardized, and affecting the quality of the tobacco products. This requires establishing a tobacco automatic grading system which, is consistent with the perception of the human vision. It can extracted tobacco-related parameters by the system, and calculated the parameters, ultimately determine the level of the tobacco based on the results of these calculations. Intelligent device has been widely used in online testing and grading screening of food, vegetables, fruits and other agricultural products, and its use not only reduce production time and labor intensity, and also improve the accuracy of online testing.This system is mainly quantitative analyse tobacco-related characteristic parameters, and to realize classify automation grading of tobacco leaves. The following aspects of study are included in this paper:1Quantization tobacco characteristic parameters. According to the standards of the national tobacco, quantify the tobacco characteristic parameters such as color, breakage rate, length and oil, it provide the basis of practice for tobacco leaves intelligent grading.2Pretreatment of grading system. Consider the impact of external light source, transmission band, and the impurity of the grading system, the collected tobacco Images require flatness correction, binarization processing, and impurities removed.3CIEDE2000tobacco color analysis. Calculate the color difference of tested tobacco leaves and all standard tobacco leaves based on CIEDE2000chrominance comparison. The grade of the current tobacco to be measured by looking for the smallest chromatic aberration and the threshold value set in the standard library. The experimental validation CIEDE2000color analysis method is better than the RGB color analysis method.4Tobacco breakage rate statistics and length measurements. By neighborhood comparison method to find damaged areas in tobacco and calculate the corresponding area of the damage; Calculate the length of tobacco leaf by the node of the test object and external rectangular bounding which made by horizontal and vertical direction of the current test tobacco.The tobacco leaf automatic grading system based on color, breakage rate, the length of the tobacco leaf built on a large number of experiments, the results show that the above method is basically consistent with the perception of the human vision, and can be more accurate identify and analyse tobacco leaf color, area, length.