基于EPO算法去除水分影响的土壤有机质高光谱估算

发布时间：2018-06-07 19:47

本文选题：土壤光谱 + 有机质　；参考：《土壤学报》2017年05期

【摘要】：野外进行土壤有机质的光谱快速预测时需考虑土壤含水量的影响。在室内设计人工加湿实验分别获取9个土壤含水量梯度(0~32%,间隔4%)的土壤光谱数据,分析土壤含水量变化对光谱的影响,再利用外部参数正交化法(external parameter orthogonalization,EPO)进行湿土光谱校正,并结合偏最小二乘回归和支持向量机回归分别建立土壤有机质预测模型。结果表明,土壤光谱反射率随着土壤含水量的增加呈非线性降低趋势,偏最小二乘回归模型的预测偏差比为1.16,模型不可用;经EPO算法校正后,各土壤含水量梯度之间的光谱差异性降低,能实现土壤有机质在不同土壤含水量梯度的有效估算,偏最小二乘回归和支持向量机回归模型的预测偏差比分别提高至1.76和2.15。研究结果可为田间快速预测土壤有机质提供必要参考。
[Abstract]:The effect of soil water content should be taken into account in rapid spectral prediction of soil organic matter in the field. The soil spectral data of 9 soil moisture gradients were obtained by artificial humidification experiment. The influence of soil moisture content on the spectrum was analyzed, and the external parameter orthogonization method was used to correct the spectrum of wet soil. Combined with partial least square regression and support vector machine regression, the prediction models of soil organic matter were established. The results show that the spectral reflectance of soil decreases with the increase of soil moisture content, and the prediction deviation ratio of partial least square regression model is 1.16, and the model is not available, which is corrected by EPO algorithm. The spectral differences among soil moisture gradients decreased, and the effective estimation of soil organic matter in different soil moisture gradients could be realized. The prediction deviation ratios of partial least square regression and support vector machine regression models were increased to 1.76 and 2.15 respectively. The results can provide a necessary reference for rapid prediction of soil organic matter in the field.
【作者单位】：华中师范大学地理过程分析与模拟湖北省重点实验室;华中师范大学城市与环境科学学院;School
【基金】：国家自然科学基金项目(41401232) 中央高校基本科研业务费专项资金项目(CCNU15A05006) 湖北省自然科学基金面上项目(2016CFB558) 华中师范大学研究生教育创新资助项目(2016CXZZ15)共同资助~~
【分类号】：S153.621

【相似文献】