摘要
通过分析土壤的主要养分含量与无人机遥感影像之间的关系,对于遥感在农业生产中的应用有重要意义。选取了昆明市云南农业大学后山试验田为研究区域,用无人机挂载多光谱相机获取研究区多光谱遥感影像。采集0~20 cm的土壤样本,并检测了土壤的理化性质及主要养分含量。通过多光谱影像不同光谱反射率及合成指数值对土壤主要养分含量进行相关性及多元逐步回归分析。结果表明:单一波段无法建立较好的回归模型,通过波段合成指数可以有效提高建模的精度。其中OSAVI、DVI、NDVI指数与K的三次方曲线的反演效果较好,R^(2)=0.641,RMSE=49.74。表明该反演模型有较高的精度和稳定性,这为遥感技术在土壤养分含量的快速测定提供了新的途径。
The relationship between the main nutrient content of the soil and UAV remote sensing images was analyzed,which is of great significance for the application of remote sensing in agricultural production.The Houshan experimental field of Yunnan agricultural university in Kunming City was selected as the research area,and a drone mounted multispectral camera was used to obtain multispectral remote sensing images of the study area.Soil samples of 0~20 cm were collected,and the soil physical and chemical properties and main nutrient contents were detected.In the experiment,the correlation and multiple stepwise regression analysis of the main nutrient content of the soil were carried out by the different spectral reflectance of the multi-spectral image and the composite index value.The results show that a good regression model cannot be established for a single band,and the accuracy of modeling can be effectively improved through the band synthesis index.Among them,the inversion effect of the cubic curve of OSAVI,DVI,NDVI index and K is better,R^(2)=0.641,RMSE=49.74.It shows that the inversion model has high accuracy and stability.It provides a new way for the rapid determination of soil nutrient content by remote sensing technology.
作者
李亚强
杨栋淏
刀剑
王建雄
LI Ya-qiang;YANG Dong-hao;DAO Jian;WANG Jian-xiong(College of Water Conservancy,Yunnan Agricultural University/Research Center of Agricultural Remote Sensing and Precision Agriculture Engineering in Yunnan University,Kunming 650201,China)
出处
《江西农业学报》
CAS
2021年第9期63-67,82,共6页
Acta Agriculturae Jiangxi
基金
云南省教育厅科学研究基金项目(2020Y0177)。
关键词
土壤主要养分
多光谱遥感
合成光谱
多元逐步回归分析
Soil main nutrients
Multispectral remote sensing
Synthetic spectrum
Multiple stepwise regression analysis
