基于支持向量机的水稻叶面积测定  

Determination of rice leaf area based on support vector machine

作　　者：孙玉婷 杨红云[2,3] 王映龙 周琼[1] 孙爱珍 杨文姬 SUN Yu-ting;YANG Hong-yun;WANG Ying-long;ZHOU Qiong;SUN Ai-zhen;YANG Wen-ji(School of Computer and Information Engineering,Jiangxi Agricultural University,Nanchang 330045,China;School of Software,Jiangxi Agricultural University,Nanchang 330045,China;Key Laboratory of Agricultural Information Technology of Colleges and Universities in Jiangxi Province,Nanchang 330045,China)

机构地区：[1]江西农业大学计算机与信息工程学院,江西南昌330045 [2]江西农业大学软件学院,江西南昌330045 [3]江西省高等学校农业信息技术重点实验室,江西南昌330045

出　　处：《江苏农业学报》2018年第5期1027-1035,共9页Jiangsu Journal of Agricultural Sciences

基　　金：国家自然科学基金项目(61562039;61363041;61462038);江西省教育厅科技项目(GJJ160374;GJJ170279)

摘　　要：本研究探讨了水稻叶长、叶宽与叶面积之间的关系。应用改进网格搜索算法参数优化后的支持向量机对计算机图像处理技术测量所得到的样本数据进行训练,以水稻叶片长度(L),叶片最大宽度(W)、叶面积回归拟合值(S)作为输入变量对水稻叶面积进行预测。结果表明,在不同的水稻品种下,相比于其他模型本研究提出的支持向量机Ⅲ模型预测结果的均方根误差和平均相对误差最低,均方根误差分别为1.882 1(两优培9),1.387 3(金优458)和1.348 2(中早35),平均相对误差分别为2.901 4%(两优培9),5.273 5%(金优458)和4.929 3%(中早35)。该模型能较真实地反映水稻叶面积的实际大小,满足农业科学研究的需求,为植物叶面积的预测提供了一种新的方法。This study explored the relationship among rice leaf length,leaf width and leaf area.The support vector machine(SVM)based on improved grid search algorithm was used to train the sample data obtained by the computer image processing technique.Rice leaf area was predicted by using the leaf length(L),leaf maximum width(W)and leaf area regression fitting value(S)as input variables.The results showed that the root mean square error and average relative error of SVMⅢmodel prediction results proposed in this study were the lowest compared with other models.The root mean square error were 1.882 1(LYP9),1.387 3(JY458)and 1.348 2(ZZ35),respectively.The average relative errors were 2.901 4%(LYP9),5.273 5%(JY458)and 4.929 3%(ZZ35),respectively.The model can reflect the actual size of rice leaf area more faithfully and meet the needs of agricultural science research,which provides a new method for plant leaf area prediction.

关 键 词：叶长 叶宽 叶面积 网格搜索算法 支持向量机 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]