西南农业学报2025，Vol.38Issue(10)：2076-2088,13.DOI:10.16213/j.cnki.scjas.2025.10.004
基于冠层结构影响的酿酒葡萄冠层叶绿素含量无人机遥感反演研究
Inversion study of chlorophyll content in canopy of wine grapes based on influence of canopy structure and soil background
摘要
Abstract
[Objective]The vegetation canopy structure and soil background significantly influence the retrieval of crop chlorophyll content.The present paper aimed to enhance the accuracy of chlorophyll content retrieval in wine grapes and thereby safeguard yield and quality of the grapes.[Method]The study focused on vineyards in the eastern foothills of Helan Mountain.Multispectral UAV data,along with leaf area in-dex(LAI)and SPAD measurements,were acquired during the flowering and fruit-setting stages.Canopy reflectance was simulated using the PROSAIL-D model.To mitigate soil background interference,the normalized difference vegetation index(NIRV)was employed for canopy effect correction,yielding the canopy scattering coefficient(CSC).Chlorophyll content estimation models were subsequently developed through vegetation index lookup tables and machine learning approaches.[Result]Correlation analysis between model-simulated vegetation indices and chlorophyll content revealed that CSC-based indices-LICI(720,750,450,840 nm),TCARI(750,720,450 nm)and EVI(720,450,750 nm)-exhibited strong correlations(|R|≥0.83)with chlorophyll levels across both growth stages.The vegetation index lookup table approach demonstrated optimal performance with EVI-TCARI combination during flowering stage(R2=0.77,RMSE=1.45),and LICI-TCARI during fruit-setting stage(R2=0.78,RMSE=1.17).Machine learning comparisons showed that PLSR and ELM models incorpora-ting CSC achieved superior accuracy:at flowering stage,ELM outperformed with R2=0.89 and RMSE=1.00(ELM＞PLSR＞ENR＞SVR);during fruit-setting,PLSR showed the best performance with R2=0.82 and RMSE=1.07(PLSR＞ELM＞SVR＞ENR).[Conclu-sion]NIR v-based canopy structure correction effectively reduces structural and soil background interference,enhancing chlorophyll content estimation accuracy.These findings establish a theoretical foundation for UAV-based vineyard monitoring and provide methodological refer-ences for rapid acquisition of grape growth parameters.关键词
无人机多光谱/PROSAIL-D/NIRV/机器学习/植被指数查找表/葡萄冠层叶绿素Key words
UAV multispectra/PROSAIL-D/NIRV/Machine learning/Vegetation index lookup table/Grape canopy chlorophyll分类
农业科技引用本文复制引用
张晓晶,潘海珠,董彦斌,张丽娟,史银龙..基于冠层结构影响的酿酒葡萄冠层叶绿素含量无人机遥感反演研究[J].西南农业学报,2025,38(10):2076-2088,13.基金项目
宁夏回族自治区研发项目(2021BEB04038) （2021BEB04038）
国家重点研发计划项目(2021YFD1900604) （2021YFD1900604）
宁夏自然科学基金项目(2021AAC03115) （2021AAC03115）
