基于光谱分析技术的苹果花钾素含量估测研究  

Estimation of Kalium Content in Apple Flowers Based on Spectral Analysis Technique

作　　者：朱西存[1,2] 赵庚星[1] 隋学艳[3] 雷彤[1] 孙顶国 

机构地区：[1]山东农业大学资源与环境学院,山东泰安271018 [2]山东农业大学农业生态与环境重点实验室,山东泰安271018 [3]山东省农业科学院,山东济南250100 [4]烟台市农业局,山东烟台264000

出　　处：《红外》2010年第8期19-23,43,共6页Infrared

基　　金：国家863计划项目(2008AA10Z203);国家级星火计划重点项目(2007EA740002);"十一五"国家科技支撑计划项目(2006BAD10A10)

摘　　要：用光谱分析技术手段探索了苹果花钾素含量的估测方法。先对苹果花的原始光谱反射率(R_i)及其8种光谱变换形式(R′_i、R″_i、1/R_i、(1/R_i)′、lg(1/R_i)、(lg(1/R_i))′、lgR_i、(lgR_i)′)与钾素含量进行相关分析,找出与钾素含量相关性较显著的光谱形式;然后,采用逐步回归分析方法,筛选敏感波长与光谱参数,建立钾素含量估测模型;进而对模型进行优选和检验,确定最佳估测模型。结果表明,苹果花光谱反射率的一阶微分R′_i与钾素含量的相关性最显著;以筛选出的敏感波长669nm和1442nm的一阶微分建立的估测模型具有较好的线性趋势;经检验,其拟合方程的R^2为0.6955,RMSE为2.7,RE%为4.6。这表明该模型对苹果花钾素含量的估测具有较好的准确度,可作为最佳估测模型。该研究为快速估测苹果花营养元素含量提供了理论依据与参考。The kalium content in apple flowers are estimated by using a spectroscopic technology. Firstly,the correlation of the original spectral reflectance（R_i）and 8 spectral transformation forms（R＇_i, R″_i,1/R_i（1/R_i）′,lg（1/R_i）,（lg（1/R_i））′,lgR_i,（lgR_i）′）of apple flowers with the kalium content are analyzed,so as to find the most significant spectral forms.Then,the stepwise regression analysis method is used to select the sensitive wavelengths and spectral parameters and establish kalium content estimation models.Finally,the best estimation models are determined through model verification.The result shows that the first derivative of spectral reflectance R′_i is most correlated with the kalium content in apple flowers.The estimation model established with the 669nm and 1442nm sensitive wavelengths of the first derivative has the best linear trend.In its fitting equation,R^2 is 0.6955,the root mean square error is 2.7 and the relative error RE%is 4.6.This indicates that the model has a better estimation accuracy for the kalium content in apple flowers and can be used as the best estimation model.This study provides the theoretical basis for the fast estimation of nutrient content in apple flowers.

关 键 词：光谱分析 苹果花 钾素含量 估测 

分 类 号：O433.4[机械工程—光学工程]