锦橙叶片叶绿素含量诊断模型—基于数字图像分析技术  

Diagnosis Model of Jincheng Orange Leaf Chlorophyll Content—Based on Digital Image Analysis Technique

作　　者：易时来[1,2] 邓烈[1,2] 何绍兰[1,2] 郑永强[1,2] 王亮[3] 赵旭阳[3] 牛廷香[3] 

机构地区：[1]西南大学柑桔研究所,重庆400712 [2]中国农科院柑桔研究所,重庆400712 [3]西南大学园艺园林学院,重庆400715

出　　处：《农机化研究》2011年第4期110-115,共6页Journal of Agricultural Mechanization Research

基　　金：重庆市自然科学基金项目(CSTC;2010BB1152);农业部科技跨越计划专项资金项目(农业部农财发[2008]42号文);公益性行业(农业)科研专项经费项目(nyhyzx07-023);科技人员服务企业行动项目(2009GJF10004);重庆市重大攻关项目(CSTC;2006)

摘　　要：以蓬安100号锦橙为试材,运用数字图像分析技术建立快速测定叶片叶绿素含量的方法。利用数码相机采集锦橙叶片图像,并运用Photoshop软件获取图像的颜色特征参数,对其进行数学变换和归一化处理后的颜色特征参数与锦橙叶片SPAD值做相关分析,建立回归模型。结果表明:颜色特征参数R/B,R/(G+B),B-R,(B-R)/(B+R),色度坐标r,r-b与SPAD值呈极显著非线性相关,并且颜色特征参数R/(G+B)和归一化坐标值r所建立的蓬安100号锦橙叶片SPAD值预测模型分别为SPAD=-1077.936[R/(G+B)]2+823.594[R/(G+B)]-74.432(R2=0.840)和SPAD=-7 883.574r2+4 715.912r-628.263(R2=0.841),其预测误差相对最小(2.12%),因此可以利用颜色特征参数R/(G+B)和色度坐标r作为基于计算机视觉的锦橙叶片SPAD值的最佳预测指标。The color characteristic parameters of the images of Peng＇ an 100 Jinchengg leaves captured with simple digital camera were acquired by the Photoshop software. Then the relationships between the characteristic color parameter processed by the mathematic transforming and normalization processed and the SPAD value of the Jincheng 100 leaves were analyzed, and the regression model were established. The results showed that the color characteristic parameter R/B, R/ （ G ＋ B）, B - R, （ B - R） / （ B ＋ R）, chromaticity coordinate r and r - b are presented a significant non - linear relativity. The SPAD value predicting model of peng＇ an 100 jincheng orange leaves were established by the color characteristic parameter R/ （ G ＋ B） and the normalized coordinates r are SPAD = - 1077. 936 [ R/（ G ＋ B ） ] ^2 ＋ 823. 594 [ R/（ G ＋ B） ] -74. 432（R^2 =0. 840）and SPAD = -7883. 574r^2 ＋4715, 912r-628. 263（R^2 =0. 841 ）, separately. The errors of prediction are relatively minimum （ 2.12 % ）. Therefore, the color characteristic parameter R/ （ G ＋ B） and chromaticity coordinate r are able to regard as best prediction index of Jincheng leaves＇s SPAD values based on computer visual.

关 键 词：锦橙叶片 数字图像处理 色彩参数 SPAD值 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] S126[自动化与计算机技术—计算机科学与技术]