摘要
为了探寻快速、准确估测土壤有机质含量的方法以推动精准农业化进程,以北疆绿洲农田灰漠土为研究对象,通过野外实地调查收集土壤样品,室内化学分析测得土壤样品有机质含量,暗室内利用SVC HR-768高光谱仪测定土壤样品光谱反射率。通过对土壤光谱反射率进行倒数、对数、一阶微分、倒数的一阶微分、对数的一阶微分变换,运用单相关分析法提取土壤光谱特征波段,采用多元逐步方法对土壤有机质含量定量反演,分析研究土壤有机质含量和室内土壤光谱的特征关系。结果表明,在波长567、1 697 nm和2 221 nm处,采用反射率对数的一阶微分建立的土壤有机质含量反演模型预测精度最高,模型决定系数达到0.82。北疆绿洲农田灰漠土土壤有机质含量高光谱反演模型的建立为土壤有机质的快速测定提供了新的途径。
This study aimed to explore a rapid and accurate method for estimating soil organic matter content and promote the process of precision agriculture. In this study, forty-six soil samples were collected from the Northern Xinjiang Oasis farmland in China. The soil organic matter of these samples was determined in laboratory. Meanwhile, the spectral reflectance of samples was measured at the indoor dark environment by the SVC HR-768 hyper-spectral spectrometer. The spectral reflectance data was transformed to several spectral indices to analyze the relationship with soil organic matter content and extract sensitive bands with correlation analysis. Quantitative inversion model of soil organic matter content was carried out by using the stepwise multiple linear regression. The results indicated that the inversion model was the best when using spectral reflectance logarithm of the first derivate for estimation at bands of 567, 1 697 nm and 2 221 nm. The coefficient of determination was 0.82. The hyper-spectral inversion model of grey desert soil organic matter content of Northern Xinjiang Oasis farmland provides a new approach for rapid soil organic matter monitoring.
作者
单海斌
蒋平安
颜安
朱磊
郭星
SHAN Hai-bin;JIANG Ping-an;YAN An;ZHU Lei;GUO Xing(College of Grassland and Environment Science , Xinjiang Agricultural University, Urumqi 830052, China;Xinjiang Key Lab of Soil and Plant Ecological Processes, Urumqi 830052, China)
出处
《农业资源与环境学报》
CAS
北大核心
2018年第3期276-282,共7页
Journal of Agricultural Resources and Environment
基金
新疆自治区重点研发计划项目(2016B063041-1)
农业部"引进国际先进农业科学技术"项目(2016-X44)
关键词
高光谱
土壤有机质
反演
模型
北疆
hyper-spectral
soil organic matter
inversion
model
Northern Xinjiang
