用高光谱数据诊断水分胁迫下棉花冠层叶片氮素状况的研究

利用ASD地物光谱仪,获取北疆棉花冠层关键生育时期的高光谱数据,应用一阶微分光谱,衍生出基于光谱位置变量的分析方法,以红边积分面积(SDr)为自变量,冠层全氮(TN)含量为因变量,做相关分析,结果表明:红边积分面积变量与冠层TN含量呈显著的相关性,相关系数是0.7394(n=40),利用构建的相关模型可以较为精确地估测棉花两个品种新陆早6号与8号冠层叶片的全氮含量,均方差(RMSE)分别为0.3859和0.4272。研究认为面积变量具有预测棉花冠层全氮含量的应用潜力。

Study on Chlorophyll and N-nutrient Contents in Cotton Canopies under Moisture Stress by Hyperspectral Remote Sensing Technology

Moisture stress in growing season of cotton can directly affect the N-nutrient absorption, growth and yield. Currently, there is no simple, rapid, effective and undamaged way for recognizing the N-nutrient content of cotton canopies. Based on the hyperspectral analysis technology and the feasibility of deriving the information of N-nutrient content of cotton canopies, in this paper, a study on the hyperspectral parameters and the total nitrogen content in cotton leaves in the main growth stages of cotton under the different irrigation conditions is carried out, which can provide the applied basis for monitoring the dynamical change of N-nutrient of cotton canopies under moisture stress by remote sensing technology.The results showed that the increase or decrease trend of the chlorophyll and total nutrient contents of two cotton varieties is same in the five main growth stages of cotton (i.e., the seedling, budding, florescence, boll and boll-opening stages) under the different irrigation conditions. The contents of both total nitrogen and chlorophyll are higher and the total area of cotton leaves increases after the florescence, and there is a significant positive correlation between the chlorophyll content and the total nitrogen content (r=0.8732, n=39). The data of chlorophyll content of cotton leaves can be used to effectively estimate the total nitrogen content by analyzing the chlorophyll content of cotton leaves in laboratory and developing the mathematical models. The hyperspectral data of cotton canopies in the main growth stages of cotton in north Xinjiang are measured by ASD FieldSpec FR, and the biochemical parameters of cotton are derived, estimated and predicted by hyperspectral remote sensing technology. The varified results of the study are as follow: (1) The analysis method based on the spectral position variables is developed, in which the integral area of red edge and the total nitrogen content of cotton canopies are regarded as the independent variable and the dependent variable respectively, the correlation between the integral area of red edge and the total nitrogen content of cotton canopies is significant, and the correlative coefficient is (0.7394) (n=40). It is considered that there is an applying potential to predict the total nitrogen content of cotton canopies with the variable of the integral area of red edge. (2) The chlorophyll content is high and the N-nutrient metabolizes violently when the moisture is sufficient, the cotton grows luxuriantly, and the red edge shifts towards blue light. It is considered that it is feasible to recognize the moisture stress of cotton by combining the red edge shifting with the change of red edge height.