Tobacco leaf intelligent grading system and method based on machine vision and hyperspectral technology

Abstract

The invention discloses a tobacco leaf intelligent grading system and method based on machine vision and hyperspectral technology, including a light box main body, an industrial CCD, a filter assembly, a warm light source group, a polyvinyl fluoride reference whiteboard and a computer, and the warm light source group is installed in the light box main body , the polyvinyl fluoride reference whiteboard is set at the bottom of the main body of the light box, the industrial CCD and the filter assembly are set at the top of the main body of the light box, and the tobacco leaves to be tested are located at the bottom of the main body of the light box. The invention uses high-speed image processing technology to process the collected tobacco leaf images, and obtains the geometric shape features, color features, texture features and internal quality features of the tobacco leaves at one time. The design structure is simple and stable, and the intelligent grading of tobacco leaves is realized, and the grading speed is very fast Fast, greatly improving the efficiency and objectivity of tobacco leaf grading, unifying the grade indicators in the tobacco leaf procurement process, improving the grade purity, providing guarantee for the better use of tobacco leaves in the cigarette industry, and at the same time improving the grading efficiency and reducing production costs.

Description

technical field [0001] The invention relates to a tobacco leaf grading system, in particular to a tobacco leaf intelligent grading system and method based on machine vision and hyperspectral technology. Background technique [0002] China is a big tobacco leaf producing country in the world. Tobacco leaf grading is a process of tobacco leaf quality appraisal and a necessary link for farmers, businessmen and industry to judge the value of tobacco leaves. So far, tobacco leaf grading is carried out manually. Due to the large differences in the classification techniques of tobacco farmers, the grades are confused and the purity is not high. This has caused disputes in the tobacco leaf purchase process, and at the same time, it has made better use of tobacco leaves of different values ​​for the industry. cause difficulty. In addition, the tobacco leaf grading process is labor-intensive and involves a lot of labor, so the grading cost is high. [0003] In recent years, many sci...

AI Technical Summary

Problems solved by technology

So far, tobacco leaf grading is carried out manually. Due to the large differences in the classification techniques of tobacco farmers, the grades are confused and the purity is not high. This has caused disputes in the tobacco leaf purchase process, and at the same time, it has made better use of tobacco leaves of different values ​​for the industry. cause difficulty

In addition, the tobacco leaf grading process is labor-intensive and involves a lot of labor, so the grading cost is high

[0003] In recent years, many scientific researchers and scientific research units have carried out research on automatic tobacco grading methods. Most of them use image processing and chromaticity theory to obtain tobacco leaf grading features for tobacco leaf grading. These features are difficult to contain the internal quality information of tobacco leaves; there are also Partially use hyperspectral imaging technology or infrared spectrum analysis to obtain the internal quality characteristics of tobacco leaves, but they cannot obtain external characteristics such as the geometric shape of tobacco leaves at the same time, and have the disadvantages of long acquisition time, large amount of data or damage to tobacco leaves; The invention patent "Method for Grading Tobacco Leaf Using Spectral and Image Features" comprehensively utilizes image processing and infrared spectrum analysis methods to obtain the external and internal characteristics of tobacco leaves, and can accurately classify tobacco leaves, but the equipment is complicated and costly. The spectral information sampling area is small, otherwise the scanning and collection time will be long, which is not conducive to the online tobacco leaf grading system

Images