基于无人机多光谱技术的田间烟叶成熟度判别研究  

Maturity Determination of Tobacco Leaves in the Field Based on UAV Multi-spectral Technology

作　　者：胡心雨 叶惠源 向东 范伟 路英健 荀拾虎 陆晓静 易克 HU Xinyu;YE Huiyuan;XIANG Dong

机构地区：[1]湖南中烟工业有限责任公司,长沙410019 [2]湖南农业大学食品科学技术学院,长沙410128 [3]湖南农业大学生物科学技术学院,长沙410128

出　　处：《智慧农业导刊》2025年第8期1-5,共5页JOURNAL OF SMART AGRICULTURE

基　　金：湖南中烟科技计划项目(KY2024XX0005)。

摘　　要：为实现基于无人机多光谱技术的田间烟叶成熟度精准判别,利用无人机搭载多光谱设备对不同时期烟田中的烟叶成熟度进行数据采集。采用斯皮尔曼相关性分析及最小冗余最大相关分析(mRMR)方法,筛选出与烟叶田间成熟度显著相关的植被指数(RBRI、NDRE、GCI、ARVI)。随后,基于筛选结果,运用随机森林方法构建烟田烟叶成熟度判别模型。研究结果显示,无人机多光谱技术能精确捕捉田间烟叶成熟度的特征差异。该判别模型的训练集和测试集准确率分别达到95.72%和93.89%。该研究有助于精准把握烟叶最佳采收时机,提升烤后烟叶品质,为大田农情监测与精准管理提供新的技术支撑。In order to accurately determine the maturity of tobacco leaves in the field based on drone multi-spectral technology,multi-spectral equipment mounted by drones was used to collect data on the maturity of tobacco leaves in tobacco fields at different periods.Spearman correlation analysis and mRMR method were used to screen out vegetation indices(RBRI,NDRE,GCI,ARVI)that were significantly related to tobacco field maturity.Subsequently,based on the screening results,a discriminant model for tobacco maturity in tobacco fields was constructed using random forest method.The research results show that drone multi-spectral technology can accurately capture the characteristics of tobacco maturity in the field.The accuracy rates of the training set and test set of the discriminant model reach 95.72%and 93.89%respectively.This research helps accurately grasp the best harvest opportunity of tobacco leaves,improves the quality of cured tobacco leaves,and provides new technical support for field agricultural monitoring and precise management.

关 键 词：无人机多光谱 植被指数 烟叶成熟度判别 相关性分析 随机森林 

分 类 号：S572[农业科学—烟草工业]