基于非接触式电导信号的土壤速效钾含量检测方法

该文针对土壤中速效钾含量采用传统测定法操作复杂、检测时效滞后的问题,建立了基于非接触式电导检测信号快速检测土壤速效钾含量的分析方法。采用高效毛细管电泳/非接触式电导检测仪获取河南潮土的非接触式电导检测信号,并使用导数法与高斯曲线拟合法相结合进行初步峰谱识别,按条件进行峰值过滤后,引入基于Levenberg-Marquardt(L-M)的高斯分峰拟合算法,实现了单峰和重叠峰的拟合计算,得到高斯峰和相应的特征参数,包括峰位、峰高、半峰宽和峰面积;最后将拟合得到的高斯峰及相应的特征参数表征原始非接触式电导检测信号离子峰谱信息结合偏最小二乘法(PLS),确定特征参数与土壤速效钾含量的关系,建立模型,实现了对土壤中速效钾含量的预测。结果表明,将基于L-M的高斯分峰拟合算法结合偏最小二乘法应用于非接触式电导检测信号测定土壤速效钾含量时具有较高精度,回归模型决定系数(R~2)为0.856 4,相对分析误差(RPD)为2.639,适用于土壤速效钾的快速检测分析。

A Detection Method for Soil Available Potassium Content Based on Capacitively Coupled Contactless Conductivity Signal

In this paper,a method based on the capacitively coupled contactless conductivity detection(C4D) signal was proposed for the rapid detection of soil available potassium content as many problems had occurred in the determination of soil available potassium content by traditonal methods,such as complex operation and long detection time.The Henan fluvo aquic soil samples were firstly characterized through high performance capillary electrophoresis with C4D device to obtain their C4D signals.Subsequently,the peak spectrums of the signals were identified by using derivative method and the Gaussian curve fitting method,then the peaks were filtered based on setting conditions.A model for single Gaussian peak was established base on Gaussian peaks,peak position,full width at half maxima and peak area of the Gaussian curve.Then Levenberg-Marquardt(L-M) algorithm was combined with the Gaussian peak fitting algorithm to optimize the characteristic parameters used to describe the information of C4D signal.The relational model between peak spectrum and soil available potassium content was established by the combination of the Gaussian characteristic parameters and partial least squares regression method(PLS),achieving the prediction of soil available potassium content.The experimental results showed that it was feasible to use the Gaussian peak fitting algorithm based on L-M algorithm combined with PLS method for C4D signal to determine soil available potassium content with high accuracy.The coefficient of determination(R2) was 0.856 4,and the residual predictive deviation(RPD) was 2.639.The method was suitable for the rapid analysis on content of soil available potassium.