无人机可见光影像对甜菜叶绿素含量的反演研究  

Research on the Inversion of Chlorophyll Content of Sugar Beet by UAV Visible Light Image

作　　者：刘智鑫 练雪萌 汪旭 吴啟贤 於丽华[1] 耿贵[1] LIU Zhi-xin;LIAN Xue-meng;WANG Xu;WU Qi-xian;YU Li-hua;GENG Gui(College of modern agriculture and ecological environment,Heilongjiang University,Harbin 150080)

机构地区：[1]黑龙江大学现代农业与生态环境学院,哈尔滨150080

出　　处：《中国甜菜糖业》2023年第4期23-31,共9页China Beet & Sugar

基　　金：国家糖料产业技术体系项目“甜菜种植制度”(CARS-170209)。

摘　　要：为了研究无人机可见光植被指数与甜菜叶绿素含量(chl)的相关性,实时监测管理甜菜生长状况,避免传统破坏性取样,指导精准施用氮肥。基于9个不同施氮水平下甜菜可见光遥感影像和田间实测冠层叶绿素含量,采用大疆精灵4无人机对甜菜生长进行监测,建立与叶绿素含量相关的支持向量机(SVM)、随机森林(RF)、岭回归(RR)机器学习模型,并利用决定系数R^(2)、均方根误差RMSE、归一化均方根误差NRMSE、平均绝对误差MAE等指标评价模型精度。结果表明:在甜菜糖分增长期,使用SVM模型的chl反演精度表现最好,R^(2)达到0.74,RMSE为0.68,MAE为0.48,NRMSE为0.16。在甜菜糖分积累期,只有SVM模型符合要求,R^(2)为0.74,RMSE为0.27,MAE为0.18,NRMSE为0.16。在甜菜整个生长期,SVM模型的chl反演精度仍然最高,R^(2)为0.76,RMSE为0.53,MAE为0.33,NRMSE为0.09。因此,无人机遥感技术在甜菜叶绿素含量估测中具有重要的理论意义,可以实现精准施氮、增产、增糖和增收。In order to study the correlation between UAV visible light vegetation index and chlorophyll content(chl)of sugar beet,real-time monitoring and management of sugar beet growth status,avoid traditional de-structive sampling,and guide precise application of nitrogen fertilizer.Based on the visible light remote sensing images of sugar beet under 9 different nitrogen application levels and the measured canopy chlorophyll content in the field,the growth of sugar beet was monitored by DJI Jingling 4 UAV,and the support vector machine(SVM),random forest(RF)and ridge regression(RR)machine learning models related to chlorophyll content were established.The accuracy of the model was evaluated by using the determination coefficient R^(2),root mean square error RMSE,normalized root mean square error NRMSE,mean absolute error MAE and other indicators.The results showed that in the sugar growth period of beet,the accuracy of chl inversion using SVM model was the best,with R^(2)of 0.74,RMSE of 0.68,MAE of 0.48 and NRMSE of 0.16.In the sugar accumulation period of sugar beet,only the SVM model met the requirements,R^(2)was 0.74,RMSE was 0.27,MAE was 0.18,and NRMSE was O0.16.In the whole growth period of sugar beet,the chl inversion accuracy of SVM model was still the highest,R^(2)was 0.76,RMSE was 0.53,MAE was 0.33,NRMSE was 0.09.Therefore,UAV remote sens-ing technology has important theoretical significance in the estimation of chlorophyll content in sugar beet,which can achieve accurate nitrogen application,yield increase,sugar increase and income increase.

关 键 词：甜菜 无人机遥感 可见光 叶绿素 氮肥 

分 类 号：S566.3[农业科学—作物学]