基于叶形分类的5种竹子叶面积预测模型研究  

Leaf Area Predicting Model for Five Bamboo Species Based on Leaf Shape Classification

作　　者：巫娟 李思卓 刘上 胡姝珍 WU Juan;LI Sizhuo;LIU Shang;HU Shuzhen(Jiangxi Provincial Academy of Water Resource Sciences,Nanchang 330029,China;School of Forestry,Jiangxi Agricultural University,Nanchang 330045,China)

机构地区：[1]江西省水利科学院,江西南昌330029 [2]江西农业大学林学院,江西南昌330045

出　　处：《生物灾害科学》2023年第1期72-79,共8页Biological Disaster Science

基　　金：江西省研究生创新专项资金项目(YC2022-S395);国家大学生创新基金训练项目(201910410018)。

摘　　要：【目的】为提高竹类植物叶面积模型的适用性,为生产实际中的叶面积测量提供快速、准确、便捷的方法。【方法】研究通过对叶面积大小悬殊的观音竹(Bambusa multiplex var.riviereorum)、唐竹(Sinobambusa tootsik)、桂竹(Phyllostachys reticulata)、中华大节竹(Indosasa sinica)和箬竹(Indocalamus tessellatus)等5种竹子按照叶形进行分类,利用长宽积和扫描实际叶面积进行叶形分类和不分类(整体)建模,并采用均方根误差、残差平方和、预测精度来检验模型的精度,并与叶面积仪测量的面积进行对比。【结果】5个竹种按叶形分类拟合的R^(2)值(观音竹0.9884、唐竹0.9879、桂竹0.9858、中华大节竹0.9935、箬竹0.9939)均大于0.98,模型预测精度(观音竹95.08%、唐竹96.68%、桂竹96.64%、中华大节竹97.51%、箬竹97.61%)均大于95%,比整体拟合准确度高;分类拟合的均方根误差(0.07~2.58)和残差平方和(0.57~600.35)均比整体拟合小。5个竹种检验结果的R2值均大于0.98,预测精度也均大于95%。【结论】研究提出的5个竹种叶面积分类拟合测算模型准确、快捷,可用于叶片面积测量,解决了竹类植物叶片面积测量难的问题,为其相关的生态学研究和经营生产提供了基础。[Objective]This study was conducted to improve the applicability of the leaf area model for bamboo and provide a fast,accurate,and convenient method for leaf area.[Method]In this study,Bambusa multiplex var.riviereorum,Sinobambusa tootsik,Phyllostachys reticulata,Indosasa sinica and Indocalamus tessellatus were selected as the research objects.Based on the classification of leaf shapes,leaf length and leaf width,a model was built to predict leaf area,taking the area measured by leaf area meter as the control.The mean root square error,residual sum of squares and predicted precision were then used to test the errors and the goodness of the model.[Result]The R^(2) values(0.9884,0.9879,0.9858,0.9935,09939)of five bamboo species were all greater than 0.98,and the predicted accuracy of the model(95.08%,96.68%,96.64%,97.51%,97.61%)was all greater than 95%.The mean root square error(0.07-2.58)and residual sum of squares(0.57-600.35)of the classification fitting were smaller than that of the overall fitting.The R2 values of five bamboo species were all greater than 0.98,and the prediction accuracy was greater than 95%.[Conclusion]Therefore,the leaf area classification and fitting calculation model of five bamboo species proposed in this paper is accurate and fast,and can be used for leaf area measurement,which solves the problem of difficult leaf area measurement of bamboo plants and provides a basis for the related ecological research and management production.

关 键 词：竹类植物 叶面积模型 叶形 叶片大小 分类拟合 

分 类 号：S795[农业科学—林木遗传育种]