品种和割龄对橡胶树叶片氮含量高光谱估算的影响

Effects of Variety and Tapping Age on Estimation of Rubber Leaf Nitrogen Content Based on Hyper-spectral

【目的】本文研究了品种和割龄对橡胶树叶片氮含量高光谱估算的影响。【方法】基于不同品种和割龄类型的橡胶树叶片样本,采用Savitzky-Golay(SG)平滑方法和偏最小二乘法(PLS)构建了橡胶树叶片氮素含量高光谱估算模型。【结果】不同品种和割龄组合下模型的预测精度从优到劣依次为:既分品种又分割龄,只分割龄不分品种,只分品种不分割龄,不分品种和割龄。【结论】品种和割龄均会影响模型精度,尤其是割龄,不同的割龄分类方法预测结果差异较大,且割龄划分并非越细越好,将橡胶树割龄分为未开割、小于10割龄和10割龄以上3种类型时预测效果最佳。

【 Objective】The effects of variety and tapping age on estimation of rubber leaf nitrogen content based on hyper-spectral was stud- ied.【 Method】 Based on rubber leaf samples of different types in varieties and tapping ages, hyper-spectral estimation models of rubber leaf nitrogen content were established by Savitzky-Golay （ SG） smoothing method and partial least squares （ PLS） ？【 Result】 The model prediction accuracy of different types in varieties and tapping ages ranking from the highest level to the lowest level were as follows, type distinguished by both varieties and tapping ages, type distinguished by tapping ages only, type distinguished by varieties only and type distinguished by neither varieties nor tapping ages.【 Conclusion】 Both varieties and tapping ages affected the precision of the mode, especially tapping ages. There were significant differences in different tapping age classification method, and it was not necessarily true that more specific division in the tapping ages would achieve better effects. In fact, the optimal prediction effects would be achieved when rubber tree tapping ages were classified as pre-tapping, tapping less than 10 years and tapping more than 10 years, respectively.