油菜氮素的多光谱图像估算模型研究

Multi-Spectral Images Estimation Models for Nitrogen Contents of Rape

【目的】充分利用可见光、近红外多谱段图像的反射强度分布信息进行油菜氮素诊断,实现对油菜氮素水平的定量估算,为油菜营养的科学管理提供理论依据和技术支撑。【方法】利用多光谱CCD获取不同氮素水平的油菜冠层图像,采用中值滤波法对图像进行预处理;通过二维最大信息熵阈值分割法对多光谱图像进行背景分割;对多光谱图像的均值和比值、差值融合特征进行提取。【结果】通过相关分析,发现ARV1、AVS560、ADV1、AVS660、g等特征在整个生长期与油菜氮素的相关性均较高,考虑到多光谱变量间存在的多重共线性影响,利用逐步回归法建立了不同生长期油菜氮素的多光谱图像特征预测模型。【结论】初步探明了不同生育时期油菜氮素的多光谱图像特征,实现了采用多光谱图像技术对油菜氮素水平的定量估算。

【Objective】Multi-spectral image analysis method was used to quantitatively analyze the rape total nitrogen content.【Method】The images of rape canopy were taken by the multi-spectral camera and were preprocessed by the median-filtering method.Two-dimensional maximum entropy segment method was used to complete background segmentation of multi-spectral images.【Result】 By extracting mean and ratio of multi-spectral images of rape canopy,it was found that the features of ARV1,AVS560,ADV1,AVS660 and g are highly correlated with nitrogen content in the whole growth period.Considering the serious multicollinearity between multi-spectral variable,the prediction model of nitrogen content of rape at different growth stages was built by stepwise regression method.【Conclusion】The reflection intensity distribution information of the visible light and the near infrared light is sufficiently utilized in this research to diagnose the nitrogen content of rape.The multi-spectral image features of the nitrogen content of rape at different growth stages were preliminarily verified.The result shows that the method of multi-spectral image analysis can be used to quantitatively analyze the rape total nitrogen content.This provides a theoretical basis and technical support for the scientific management of rape nutrition.