Tobacco frame tobacco leaf grade determination method based on deep learning

Abstract

The invention provides a tobacco frame tobacco leaf grade determination method based on deep learning. The method comprises the following steps: 1, selecting a plurality of different grades of tobacco leaves as samples, and collecting tobacco leaf images of the samples in a tobacco frame; 2, classifying and storing the acquired tobacco leaf images according to different grades; 3, constructing a prediction model, preprocessing the classified tobacco leaf images, inputting the preprocessed tobacco leaf images into the prediction model, and training the preprocessed tobacco leaf images to obtain a trained prediction model; 4, determining the grade of the tobacco leaves in the to-be-detected tobacco frame, firstly collecting the tobacco leaf image of the tobacco leaves in the to-be-detected tobacco frame, carrying out the same preprocessing as the step 3 on the collected image, and then predicting by utilizing the trained prediction model to obtain a tobacco leaf grade determination result; and 5, outputting a judgment result. The method is used for judging the grade of the tobacco leaves, reduces the manual workload and improves the production efficiency.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to a method for judging the grade of tobacco frame tobacco leaves based on deep learning. Background technique [0002] Tobacco leaf is an important raw material in tobacco production, and its quality grade will directly affect the quality of tobacco products, and the tobacco leaves selected for different tobacco products are also different. Due to the influence of other factors such as natural growth, different tobacco leaves often have more or less differences in quality. If the tobacco leaves mixed together without modulation are sold, it does not meet the requirements of various tobacco products, and the tobacco products will lose their use value. [0003] Tobacco leaf grading in the traditional method mainly relies on traditional methods such as seeing, touching, and smelling by experts or classifiers to determine the quality of tobacco leaves. [0004] Th...

AI Technical Summary

Problems solved by technology

[0004] The current method is to analyze the appearance characteristic parameters of tobacco leaves, such as color, oil content, maturity, disability, etc. After digitalizing and quantifying these indicators, automatic grading can be realized. However, the digitization of parameter indicators in this method is still Manually, the efficiency is relatively low, and the results are often irregular

Images