柑橘园紫色土Fe/Mn/Zn含量近红外光谱监测模型研究

 【目的】探索建立基于近红外光谱技术的土壤微量元素监测技术。【方法】采集三峡库区（重庆）主要加工甜橙基地果园背景土壤样品168个，随机选取100个作为建模样本，其余为检验样本；测定所有样本的近红外反射光谱和土壤Fe、Mn、Zn全含量；运用最佳光谱预处理方法和偏最小二乘法（partial least square method, PLS）及内部交叉验证方法建立校正模型，并进行模型精度检验。【结果】变量标准化（standard normal variables，SNV）为土壤Fe、Mn、Zn含量近红外光谱预测的最佳光谱预处理方法；运用SNV光谱预处理和偏最小二乘法（PLS）及内部交叉验证法建立的土壤Fe、Mn、Zn含量校正模型，95%置信区间内的预测精度分别为92.65%、95.59%和95.59%。【结论】利用近红外反射光谱技术进行土壤Fe、Mn、Zn含量检测可行且精度较高。

The Near Infrared Spectrum Based Models for Monitoring Fe,Mn,and Zn Content of the Purple Soil in Citrus Orchard

【Objective】The objective of the study is to explore and establish the monitoring approach of purple soil microelements based on near infrared spectroscopy technology. 【Method】There were 168 samples of background soil of main sweet orange orchards in Three Gorges Reservoir Area were collected. After the sample processing, 100 samples were selected as calibration samples randomly, others for test samples. The quantity of near infrared spectrum and total Fe, Mn, Zn content of purple soil were determined. By using the optimum spectrum pretreatment and partial least square method(PLS) along with internal cross- validation the calibration models were established, and the model accuracy was tested.【Result】The best spectrum pretreatment for prediction of total Fe, Mn, Zn contents of purple soil is standardization of variables by using near infrared spectroscopy. The calibration models of Fe, Mn, Zn concentration of purple soil established by using the spectrum pretreatment, PLS and internal cross-validation, the precisions under 95% confidence interval were 92.65%, 95.59% and 95.59%, respectively.【Conclusion】It has a certain feasibility by using the near-infrared spectroscopy to predict total Fe, Mn, Zn concentrations of purple soil.