论文部分内容阅读
应用高光谱技术探讨土壤有机质含量定量估测方法,对发展精细农业具有重要意义。本文利用陕西省横山县的实测数据,采用对数的一阶微分变换方法对土样的高光谱数据进行处理,分别采用线性回归分析法、BP神经网络法、模糊识别法建立高光谱土壤有机质含量估测模型,并对比分析其精度,确定最优的光谱反演模型。实验结果表明:模糊识别模型的决定系数达到0.973,RMSE为0.0468%;比线性模型和BP神经网络模型精度都高。研究表明,土壤有机质光谱反演不仅要重视机理研究,同时要加强光谱反演建模方法创新。
The application of hyperspectral techniques to the quantitative estimation of soil organic matter content is of great significance to the development of fine agriculture. In this paper, we use the first-order differential transformation method of logarithm to deal with the hyperspectral data in Hengshan County, Shaanxi Province. The linear regression analysis, BP neural network and fuzzy recognition are used to establish the hyperspectral soil organic matter content Estimate the model, and compare its accuracy, to determine the optimal spectral inversion model. The experimental results show that the decision coefficient of the fuzzy recognition model reaches 0.973 and the RMSE is 0.0468%. The accuracy of both the linear model and the BP neural network model is high. The research shows that the inversion of soil organic matter spectra should not only pay attention to the mechanism research, but also strengthen the innovation of spectral inversion modeling method.