摘要
在玉米追氮关键期对其氮含量进行准确检测,是玉米减量施肥的重要举措。传统的全波段光谱分析法数据量大,氮含量检测数学模型建立困难。选取能够反映玉米叶片氮含量水平的特征波长,能够极大减小建模难度。针对玉米生长对氮含量特征波长变化影响,提出玉米追肥关键期叶片氮含量检测方法。将标准归一化、卷积平滑处理算法与连续投影算法相结合,提取了玉米拔节期、大喇叭口期及抽雄期3个追氮关键期的叶片氮含量光谱特征波长,为玉米追氮光谱检测模型建立提供了依据,并通过多元线性回归建立玉米叶片氮含量光谱模型,极大降低了算法复杂度。通过对比分析表明,采用本文提取的方法得到的特征波长,具有准确性强,均方根误差小,运算复杂度低的优势。
Nitrogen is a vital element in corn’s growth. It is important to detect the nitrogen content accurately for reducing fertilization of corn in the critical period of maize nitrogen recovery. The spectral analysis can accurately detect nitrogen content in maize leaves, but the large amount of data in full band spectral analysis makes it difficult to establish a mathematical model for nitrogen content detection. The difficulty of modeling can be greatly reduced by selecting characteristic wavelengths which can reflect the nitrogen content level of maize leaves. However, traditional methods often extract the characteristic wavelength of a certain period as the basis for dimension reduction, which does not fully consider the influence of the change on characteristic wavelength of nitrogen content of corn growth. In this study, we extract the spectral characteristic wavelengths of nitrogen content in leaves of maize at three critical stages of nitrogen topdressing, namely jointing stage, bell mouth stage and heading stage, with the methods of standard normalization, convolution smoothing algorithm and continuous projection algorithm, which provide a basis for the establishment of the spectral detection model of nitrogen topdressing in maize. The spectral model proposed in this paper can greatly reduce the complexity of the algorithm. Through the comparison analysis of multivariate linear modeling, the feature wavelength extracted by this method has high accuracy, small root mean square error and low computational complexity.
作者
刘丹
马璐萍
李建昌
孙磊
赵建国
郝建军
LIU Dan;MA Luping;LI Jianchang;SUN Lei;ZHAO Jianguo;HAO Jianjun(College of Mechanical and Electrical Engineering,Hebei Agricultural University,Baoding 071001,China)
出处
《河北农业大学学报》
CAS
CSCD
北大核心
2020年第3期102-107,共6页
Journal of Hebei Agricultural University
基金
国家重点研发计划(2018YFD0200607)
河北省高等学校科学技术研究项目(QN202044).
关键词
氮含量
特征波长
光谱
玉米叶片
nitrogen content
characteristic wavelength
spectrum
maize leaves
