摘 要:针对名优茶嫩芽自动采摘问题,采用 SVM 学习算法实现对名优茶嫩芽图像自动分割。通过提取嫩芽像素点与背景像素点的 RGB 及(R-B)特征,将 4 个特征按重要性组合为 3 个特征组,分别是 RGB 特征组、RGB+(R-B)特征组和 G+(R-B)特征组,利用 3 个特征组分别构建 SVM 嫩芽分割模型。在收集的多幅图像上的实验表明 G+(R-B)特征组构建的分割模型分割得到的嫩芽图像较为完整,且用时均低于 0.5 s,满足名优茶嫩芽自动采摘的要求。
关键词:自动采摘;SVM;图像分割
DOI:10.19850/j.cnki.2096-4706.2021.02.022
中图分类号:TP391 文献标识码: A 文章编号:2096-4706(2021)02-0089-04
Automatic Segmentation Method of Famous Tea Buds Image Based on SVM
CHEN Miaoting,YANG Guanglei,QIN Pengtao
(Qingdao University of Science and Technology,Weifang 261500,China)
Abstract:Aiming at the problem of automatic picking of famous tea buds,the SVM learning algorithm is used to realize the automatic segmentation of the image of famous tea buds. By extracting the RGB and(R-B)features of the bud pixels and the background pixels,the four features are combined into three feature groups according to their importance. They are RGB feature group,RGB+(R-B) feature group and G+(R-B)feature group. Three feature groups were used to construct SVM buds segmentation models. Experiments on several collected images show that the segmentation model constructed by G+(R-B)feature group can obtain relatively complete bud images,and the segmentation time is less than 0.5 s,which meets the requirements of automatic picking of famous tea buds.
Keywords:automatic picking;SVM;image segmentation
参考文献:
[1] 章毓晋 . 图像分割 [M]. 北京:科学出版社,2001:78-90.
[2] GONZALEZ R C. Digital Image Processing [M].3rd ed. Upper Saddle River:Prentice Hall,2008.
[3] 周莉莉,姜枫 . 图像分割方法综述研究 [J]. 计算机应用研究,2017,34(7):1921-1928.
[4] 韦波 . 基于颜色和形状的茶叶计算机识别研究 [J]. 福建茶叶,2016,38(3):16-17.
[5] 唐仙,吴雪梅,张富贵,等 . 基于阈值分割法的茶叶嫩芽识别研究 [J]. 农业装备技术,2013,39(6):10-14.
[6] 汪建 . 结合颜色和区域生长的茶叶图像分割算法研究 [J].茶叶科学,2011,31(1):72-77.
[7] 薛志东,王燕,李利军 .SVM 图像分割方法的研究 [J].微计算机信息,2007(24):306-308.
[8] 胡涛,胡军,郭杭 . 基于 SVM 邻域学习的视频目标检测方法 [J]. 现代电子技术,2017,40(14):95-98.
[9] 韦佳佳 . 名优茶机械化采摘中嫩芽识别方法的研究 [D].南京:南京林业大学,2012.
[10] 邓乃扬,田英杰 . 支持向量机:理论算法与拓展 [M]. 北京:科学出版社,2009:10-35.
[11] BURGES C J C.A Tutorial on Support Vector Machines for Pattern Recognition [J].Data Mining and Knowledge Discovery,1998,2:121-167.
[12] 袁志华,杨百龙,赵文强,等 . 基于支持向量机的 NSCT 域自适应图像水印算法 [J]. 计算机应用研究,2018,35(6):793-
1796.
[13] CHANG C C,LIN C J. LIBSVM:A library for support vector machines [J/OL].ACM Transactions on Intelligent Systems and
Technology,2011,2(3):[2020-11-12].https://doi.org/10.1145/19611 89.1961199.
作者简介:陈妙婷(1993—),女,汉族,河北衡水人,助教,硕士,研究方向:人工智能与图像处理;
通讯作者:杨广蕾(1993—),女,汉族,山东淄博人,助教,硕士,研究方向:超硬磨料钎焊。