草地学报 ›› 2013, Vol. 21 ›› Issue (5): 902-905,920.DOI: 10.11733/j.issn.1007-0435.2013.05.010

• 研究论文 • 上一篇    下一篇

紫花苜蓿植株含氮量的高光谱估测技术研究

付彦博, 范燕敏, 盛建东, 武红旗, 李宁, 朱高飞, 刘焕鲜, 龚双凤   

  1. 新疆农业大学草业与环境科学学院, 新疆 乌鲁木齐 830052
  • 收稿日期:2013-03-18 修回日期:2013-04-28 出版日期:2013-10-15 发布日期:2013-10-30
  • 通讯作者: 盛建东
  • 作者简介:付彦博 (1986- ),男,河南驻马店人,硕士研究生,主要从事土壤与植物营养研究,E-mail:fyb6915028@163.com
  • 基金资助:

    新疆维吾尔自治区科技支疆项目(201291142);国家自然科学基金(31160113)资助

Hyperspectral Estimation Technique of Plant Nitrogen Content of Alfalfa

FU Yan-bo, FAN Yan-min, SHENG Jian-dong, WU Hong-qi, LI Ning, ZHU Gao-fei, LIU Huan-xian, GONG Shuang-feng   

  1. Xinjiang Agricultural University, College of Grassland and Environmental Sciences, Urumqi, Xinjiang 830052, China
  • Received:2013-03-18 Revised:2013-04-28 Online:2013-10-15 Published:2013-10-30

摘要: 通过研究不同施氮水平下各生育期紫花苜蓿(Medicago sativa L.)冠层光谱反射率,对不同施氮水平下苜蓿地上部器官氮含量与可见光波段(510, 560, 610, 680, 710, 760 nm)和近红外波段(810, 870, 950 nm)的植被指数(RVI, DVI, NDVI和RDVI)进行相关性分析,确立了苜蓿现蕾期叶、蕾、植株的光谱预测模型。 结果表明:760, 810和870 nm构建的植被指数与苜蓿地上部器官氮含量的相关性较好;通过筛选建立的植株氮含量与差值植被指数DVI(760, 810)光谱预测模型最优,调整R2和相对平均误差分别为0.919和9.1%。所建立的紫花苜蓿氮含量光谱预测模型可为新疆苜蓿营养诊断与施肥提供决策依据。

关键词: 紫花苜蓿, 氮肥, 冠层光谱反射率, 植被指数, 模型

Abstract: Spectral reflectance characteristics of alfalfa (Medicago sativa L.) canopy under different nitrogen levels were studied in this research. Relationships between the nitrogen contents of aboveground organs and the vegetation indices (RVI, DVI, NDVI and RDVI) of visible wavelengths (510 nm, 560 nm, 610 nm, 680 nm, 710 nm, and 760 nm), and near infrared wavelengths (810 nm, 870 nm, 950 nm) were analyzed. The spectral prediction model of nitrogen contents of alfalfa leaves, buds and plants were established respectively in budding period. Results showed that vegetation index established by 760 nm, 810 nm and 870 nm had better correlation with the nitrogen contents of aboveground organs. The optimal spectral prediction model was determined by comparison, which was vegetation index DVI(760, 810) prediction model, and adjusting R2 was 0.919, average relative error was 9.1%. Therefore,spectral prediction models established for evaluating the nitrogen contents of alfalfa in this research might provide a basis of decision making for the nutritional diagnosis and fertilization of alfalfa in Xinjiang.

Key words: Alfalfa, Nitrogen, Canopy spectral reflectance, Vegetation index, Model

中图分类号: