自然环境下茶树嫩梢识别方法研究.pdfVIP

茶叶科学 20 12，32 （5 ）：377~381 Journal of Tea Science 投稿平台： 自然环境下茶树嫩梢识别方法研究 1 1* 1 1 2 1 韦佳佳 ，陈勇 ，金小俊 ，郑加强 ，石元值 ，张浩 1. 南京林业大学机械电子工程学院，江苏 南京 210037 ；2. 中国农业科学院茶叶研究所，浙江 杭州 310008 摘要：嫩梢识别是实现名优茶智能采摘的前提。本文以茶树嫩梢为研究对象，基于色彩因子开展了自然环境下 嫩梢识别研究，提出了采用RGB 空间的R-B 、YIQ 空间的I 、Lab 空间的b 、H SI 空间的S，以及YCrCb 空间 的Cb 5 种色彩因子进行图像灰度化，并选择合适的方法进行图像阈值分割，最后采用中值滤波的方法消除噪 声。试验结果表明，这些方法都能够在自然环境下有效地区分嫩梢和背景，为后续名优茶智能化采茶机的研究 打下理论基础。 关键词：名优茶；嫩梢识别；色彩因子；阈值分割 中图分类号：S571.1 文献标识码：A 文章编号：1000-369X （2012 ）05-377-05 Researches on Tender Tea Shoots Identification under Natural Conditions 1 1* 1 1 2 1 WEI Jia-jia , CHEN Yong , JIN Xiao-jun , ZHENG Jia-qiang , SHI Yuan-zhi , ZHANG Hao 1. College of Electronic and Mechanical Engineering, Nanjing Forestry University, Jiangsu 210037, China ; 2 . Tea Research Institute, Chinese Academy of Agricultural Sciences, Hangzhou 310008, China Abstract : Identification of the tender tea shoots is the key step towards the intelligent tea harvesting. This paper presents several methods to recognize the tender tea shoots for high-quality tea production. Gray images were obtained by five color indices, which were R-B, I, b, S and Cb in RGB, YIQ, Lab , HSI and YCrCb color spaces. Then suitable threshold methods were applied to segment image, finally the median filter was used to eliminate noise s. The results indicate that these methods were particularl y effective for tender tea shoots identification under their natural conditions. The proposed method can be used for future intelligent tea harvest development. Keywords : high-quality tea, tender tea shoots