Tobacco leaf grading method based on reflection, perspective and microscopic images

Abstract

The invention discloses a tobacco leaf grading method based on reflection, perspective and microscopic images. The method provided by the invention comprises the following steps: (1) acquiring a high-definition color reflection image, a high-definition perspective image and a high-definition microscopic image of tobacco leaf at real time and transmitting the collected images to an industrial control computer; (2) preprocessing the images to quickly position the edge of tobacco leaf and segmenting tobacco leaf out of a background area to record as a tobacco image; (3) extracting features of the image, wherein the features include shape feature, color feature and texture feature; and (4) analyzing the above features of tobacco leaf by an Adaboost sorting algorithm, and determining the grade of tobacco leaf through part, color and grade classification of tobacco leaf. Through a device for full-automatic machine grading, human factor interference can be avoided, and labor cost is greatly reduced. Thus, huge hidden costs of training, management and the like caused by manual grading are completely avoided.

Description

technical field [0001] The invention relates to a method for grading tobacco leaves, in particular to a method for grading tobacco leaves based on reflection, perspective and microscopic images, which uses a pattern recognition algorithm to analyze and process the tobacco leaves. Background technique [0002] Tobacco leaf grading is to divide tobacco leaves into different grades according to their quality characteristics and degree of pros and cons, so that each grade of tobacco leaves has relatively consistent quality characteristics and quality levels for the cigarette industry to choose and use. my country officially promulgated the "National Standards for Flue-cured Tobacco of the People's Republic of China" in 1992, pointing out that tobacco leaves can be divided into 42 grades, and the grades of tobacco leaves are composed of parts, colors and grades; , the grades of tobacco leaves in different regions and years are different, and the grades of tobacco leaves are closel...

AI Technical Summary

Problems solved by technology

This method requires a large number of operators and needs to be trained. In the actual implementation process, the effect is not good

There are unfavorable factors such as large personnel mobility, large floor area, poor production environment, uneven experience and ability of operators, and high enterprise management costs.

Specifically, the quality of finished tobacco leaves is greatly affected by artificial factors, and the quality deviation of tobacco leaves after grading is relatively large.

The current tobacco leaf grading mode is not conducive to the rational use of tobacco leaf resources, resulting in large fluctuations in the quality of tobacco leaf processing, and has affected the quality of cigarette processing

[0004] The Chinese invention patent with the application number 201110004173.7 discloses a method for classifying tobacco leaves using spectral and image features, but it is not suitable for industrial needs. For example, the patent uses a neural network to delete features, which will reduce the features. The classification is more complicated, and the reduction of features may lead to inaccurate classification, which greatly reduces the classification ability of the network; when classifying tobacco leaves, several classification algorithms are not scientific, which points out that the network structure and the number of nodes of the neural network algorithm are selected It is very important in classification, because the classification of tobacco leaves in different years and different origins may vary greatly, and the network structure cannot meet various situations; genetic algorithms are usually used to optimize parameters, and are not suitable for classification; using the nearest neighbor algorithm is too computationally intensive, It is necessary to calculate the distance between the current sample and each sample; the clustering algorithm is not applicable when there are many levels, the cluster center is difficult to determine, and the amount of calculation is huge

Images