土壤含水率对近红外传感器标定模型的响应研究  

Responses of soil moisture content measured by near-infrared sensor to calibration models

作　　者：殷哲[1] 雷廷武[1] 晏清洪[1] 陈展鹏[1] 董月群[1] 庄晓晖[1] 

机构地区：[1]中国农业大学水利与土木工程学院,北京10083

出　　处：《中国农业大学学报》2013年第6期102-106,共5页Journal of China Agricultural University

基　　金：中国农业大学研究生科研创新专项项目(KYCX2010101)

摘　　要：为提高近红外传感器测量的准确度,进一步理解不同标定模型对土壤含水率测量精度的影响。利用土壤表面的近红外反射光强来预测土壤含水率,通过归一化处理将反射光强转化为相对吸收深度和相对反射率,采用2种标定方法,分别建立土壤含水率与相对吸收深度之间及土壤含水率与相对反射率之间的线性模型与非线性模型。选取我国东北地区的黑土进行标定,并用独立的试验数据对模型进行检验。结果表明,吸收深度法的线性和非线性模型的预测值和实测值符合度较好。反射率法的线性模型和非线性模型对土壤的含水率预测均方根误差(RMSE)分别为2.89%和2.95%,相对吸收深度法非线性模型的RMSE值明显大于其他3种模型,预测准确度最低。说明不同标定方法会影响土壤含水率的预测结果。4种模型的预测精度能够满足测量要求。This paper is concerned on To improving the accuracy of near-infrared soil moisture sensor and further understanding the effect of different calibration methods on prediction result of soil moisture,An approach is presented to estimate surface soil moisture from reflectance data in the wavelength of 1 940 and 1 800 nm. With two calibration methods, the reflectance values were normalized by that of maximum. Linear and nonlinear models between soil moisture content and relative absorption band depth and linear and non-linear models between soil moisture content and relative reflectance respectively were determined through calibration. The experiments were made with black soil samples to supply data for calibration of the models and . The independent data sets were used for validation of the models. The prediction results indicated high precision for absorption depth models（linear and non-liner ones）, with the root mean squared errors（RMSE） of 2.88% and 5.83% respectively. The linear and non-linear models of reflectance method had RMSE of 2.89% and 2.95%. But the accuracy of non-linear model of reflectance method was highest compared with other three models. The validation results indicated the effect of calibration method on accuracy of soil moisture prediction. All of these four models could meet the requirement of soil moisture measurement.

关 键 词：近红外 土壤含水率 反射率 吸收深度 模型 

分 类 号：S152.7[农业科学—土壤学]