去除土壤水分对高光谱估算土壤有机质含量的影响

发布时间：2018-07-18 11:43

【摘要】：土壤高光谱技术具有方便快捷、无破坏、成本低等优点,已被广泛应用于估算土壤有机质含量(SOMC)。然而,野外测量的土壤高光谱数据因受外部环境因素(土壤湿度、温度、表面粗糙度等)干扰,导致SOMC估算模型适用性有待提升。土壤含水率(SMC)是影响野外测量高光谱的最主要的障碍因素之一,它的变化严重影响可见-近红外(Vis-NIR)光谱反射率的观测结果。因此,消除SMC对高光谱数据的干扰是提高土壤高光谱估算SOMC模型预测精度的关键环节。以江汉平原潜江市潮土样本为研究对象,在室内人工加湿土样,分别获取6个SMC水平的土壤高光谱数据,采用标准正态变换(SNV)对光谱数据进行预处理,基于外部参数正交化法(EPO)去除土壤水分对高光谱的影响,利用偏最小二乘方法(PLSR)建立并对比EPO处理前、后不同SMC水平SOMC反演模型。结果表明,土壤水分对Vis-NIR光谱反射率有显著的影响,掩盖了SOMC的光谱吸收特征;EPO处理前不同SMC水平的光谱曲线之间的差异较为明显,而EPO处理后的各SMC水平的光谱曲线形态基本相似;采用EPO处理后的土壤高光谱数据建立SOMC估算模型,预测集的R2p,RPD分别为0.84和2.50,其精度与EPO处理前所建模型相比有较大提升,表明EPO算法可以有效去除土壤水分的影响,从而提升SOMC的估算精度。对定向去除外部环境参数对土壤高光谱影响进行了实证,为完善野外原位获取SOMC信息技术提供理论基础。
[Abstract]:Soil hyperspectral technology has been widely used to estimate soil organic matter (SOMC) because of its advantages of convenience, no damage and low cost. However, the soil hyperspectral data measured in the field are disturbed by external environmental factors (soil moisture, temperature, surface roughness, etc.), so the applicability of SOMC estimation model needs to be improved. Soil moisture content (SMC) is one of the most important factors affecting the measurement of hyperspectral spectrum in the field. The change of soil moisture content (SMC) seriously affects the observational results of visible and near infrared (Vis-NIR) spectral reflectance. Therefore, eliminating the interference of SMC to the hyperspectral data is the key to improve the prediction accuracy of SOMC model in soil hyperspectral estimation. Six soil hyperspectral data at SMC level were obtained by artificial humidification of soil samples from Qianjiang City, Jianghan Plain. Standard normal transform (SNV) was used to preprocess the spectral data. Based on the effect of soil moisture removal by external parameter orthogonalization (EPO), the partial least square (PLSR) method was used to establish and compare the SOMC inversion models at different SMC levels before and after EPO treatment. The results showed that soil moisture had a significant effect on the spectral reflectance of Vis-NIR, which masked the spectral absorption characteristics of SOMC and the difference between the spectral curves of different SMC levels before EPO treatment. The spectral curves of each SMC level after EPO treatment were similar, and the estimation model of SOMC was established by using the soil hyperspectral data after EPO treatment, and the R2pN RPD of the predicted set was 0.84 and 2.50, respectively, which was higher than that of the model established before EPO treatment. It is shown that EPO algorithm can effectively remove the influence of soil moisture and improve the estimation accuracy of SOMC. The effects of directional removal of external environmental parameters on soil hyperspectral were demonstrated, which provided a theoretical basis for the improvement of in situ SOMC information technology 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(41401232,41271534) 中央高校基本科研业务费专项资金项目(CCNU15A05006,CCNU15ZD001)资助
【分类号】：S153

【相似文献】