基于多光谱成像的番茄叶片叶绿素含量预测建模方法研究

传统的光谱分析技术预测植物的叶绿素含量的精度较低,而基于3CCD的多光谱摄像机的叶绿素预测研究存在其摄像机本身成本昂贵和无法调整的波长通道数等局限性.文章提出了基于多光谱图像技术利用敏感波长(532,610和700 nm)下番茄叶片的灰度值来预测其叶绿素含量的研究方法.利用多元线性回归分析、主成分分析和偏最小二乘回归分析等方法建立了预测模型,取得了较好的预测效果,其相关系数R2c与R2v均达到了0.9左右.表明该方法用于番笳叶绿素的预测是有效和可行的,也为作物的长势检测仪器的开发奠定了基础.

Research on predicting modeling for chlorophyll contents of greenhouse tomato leaves based on multi-spectral imaging

Traditional spectrum analysis technology has low accuracy for forecasting chlorophyll content of plants. Research based on 3CCD camera has the limitations of high cost and the number of sensitive wavelengths not adjustable. The present paper develops a new approach to forecasting the chlorophyll content of tomato leaves by the image gray value of the selected sensitive wavelengths (532, 610 and 700 nm). Three common methods such as multi-linear regression, principal component analysis and partial least square regression were employed in forecast modeling, the good results were obtained, and both Rc2 and Rv2 reached about 0.9. The method has proven effective and feasible for prediction of chlorophyll contents of tomato leaves, which also lays the foundation for the development of testing instruments for the growing of crops.