高级检索
    章春娥, 魏扬帆, 王岩松. 烟叶自动图像分级技术研究综述[J]. 中国烟草科学, 2014, 35(4): 103-108. DOI: 10.13496/j.issn.1007-5119.2014.04.020
    引用本文: 章春娥, 魏扬帆, 王岩松. 烟叶自动图像分级技术研究综述[J]. 中国烟草科学, 2014, 35(4): 103-108. DOI: 10.13496/j.issn.1007-5119.2014.04.020
    ZHANG Chun'e, WEI Yangfan, WANG Yansong. A Survey on Automatic Tobacco Image Grading Techniques[J]. CHINESE TOBACCO SCIENCE, 2014, 35(4): 103-108. DOI: 10.13496/j.issn.1007-5119.2014.04.020
    Citation: ZHANG Chun'e, WEI Yangfan, WANG Yansong. A Survey on Automatic Tobacco Image Grading Techniques[J]. CHINESE TOBACCO SCIENCE, 2014, 35(4): 103-108. DOI: 10.13496/j.issn.1007-5119.2014.04.020

    烟叶自动图像分级技术研究综述

    A Survey on Automatic Tobacco Image Grading Techniques

    • 摘要: 在烟叶自动分级的过程中,提取烟叶特征并且进行分级是至关重要的步骤。阐明了烟叶分级的现行国家标准,分析总结了常用的颜色、形状、纹理等特征分析手段,包括一些较为新颖的方法。对多种方法综合对比分析,结合我国烟草行业实际背景与现实环境,从简捷、经济、高效、可靠四方面考虑,提出在标准D50照明光源条件下,提取现实中可用的叶长、面积等图像特征,选择适合判别的均匀颜色模型对烟叶进行分级判别。

       

      Abstract: To obtain the characteristics that can be used for grading tobacco leaves is a crucial step in the process of an automatic tobacco image grading system. This article first expounds the current national grading standard of tobacco leaves, and then explores the feasible methods which is suitable for industrial environment after analyzing techniques using the characteristics of the color, shape, texture and other technical measure, with considering of simplicity, economy, efficiency, and reliability. The method solved the problems in LED light, image feature extracting and choice of well distributed color model.

       

    /

    返回文章
    返回