高光谱图像技术检测黄瓜叶片的叶绿素叶面分布  

Measurement of Chlorophyll Distribution in Cucumber Leaves Based on Hyper-spectral Imaging Technique

作　　者：石吉勇[1] 邹小波[1] 赵杰文[1] 殷晓平[1] 

机构地区：[1]江苏大学食品与生物工程学院,镇江212013

出　　处：《分析化学》2011年第2期243-247,共5页Chinese Journal of Analytical Chemistry

基　　金：国家863项目(No.2008AA10Z208);国家自然科学基金(No.60901079);江苏省六大人才高峰;青蓝工程项目;国家博士后基金;优秀博士论文基金;江苏省普通高校研究生科研创新计划(No.CX10B_2772)资助

摘　　要：以黄瓜叶片为材料,利用高光谱图像技术结合独立分量法(ICA),研究了叶绿素浓度叶面分布的快速、无损检测方法。用高光谱相机采集了80片黄瓜叶子的高光谱图像(408～1117 nm),利用ICA方法提取了高光谱图像的8个独立分量信号,通过逐步线性回归(SMLR)优选出第1、第2和第5个ICA信号,并在此基础上建立了叶绿素浓度回归建模(R=0.904,RMSE=0.221)。此模型表征了高光谱图像信号同叶绿素浓度的线性映射关系。提取黄瓜叶高光谱图像所有像素点的第1、第2和第5个ICA信号,代入建立的叶绿素浓度模型中,快速计算出所有像素点对应的叶绿素浓度,并用叶绿素浓度值代替对应像素的图像信号,最终得到了黄瓜叶片的叶绿素浓度分布图。研究表明,利用高光谱图像技术快速、无损检测叶片叶绿素浓度及其叶面分布是可行的。Chlorophyll distribution in cucumber leaves was non-destructively and rapidly measured by Independent component analysis（ICA） combined with hyper-spectral images.Eighty hyper-spectral images of cucumber leaves were collected by hyper-spectral camera（408-1117 nm）.Eight independent signals were computed by ICA,and the first,the second,the fifth ICA signals were used to build step multiple linear regression（SMLR） model（R=0.904,RMSE=0.221）.Then the first,the second,the fifth ICA signals of every pixel in the hyper-spectral image were computed and chlorophyll content of every pixel was got according to the multiple linear regression models.Finally,the chlorophyll distribution map was estimated.Overall results sufficiently demonstrated that the hyper-spectral imaging technology can be used to measure the chlorophyll concentration and distribution in cucumber leaves.

关 键 词：叶绿素分布 高光谱图像 独立分量分析 黄瓜叶片 

分 类 号：Q946[生物学—植物学]