In order to improve the prediction accuracy soil texture based on VIS-NIR spectra this paper introduces the orthogonal signal correction (OSC) spectra pretreatment method.Separate partial least squares regression (PLSR) model was established using the original spectrum,derivative analysis spectra,and OSC processing spectra respectively.The results showed that OSC-PLSR model validation's accuracy is higher than that of the other two models.The RMSEp of prediction from the OSC-PLSR models of content of sand and clay separates were 5.94 and 1.25,which were 22.22% and 9.42% lower than of the PLSR model respectively.The results of this research showed OSC algorithm can effectively eliminate the influence of unrelated factors and improve the accuracy of prediction when using Vis-NIR spectra to predict soil texture.%为提高基于VIS-NIR光谱的土壤质地预测精度,引入了正交信号校正(OSC)光谱预处理算法.分别用原始光谱、微分处理、OSC处理光谱,建立偏最小二乘回归(PLSR)模型.结果表明,OSC-PLSR模型验证精度高于其他两种方法所建模型,砂粒含量OSC-PLSR模型的RMSEp为5.94,粘粒含量OSC-PLSR模型RMSEp为1.25,相比PLSR模型,分别降低22.22%和9.42%.OSC算法在土壤质地的VIS-NIR反演中能有效消除不相关因素的影响,提高模型预测精度.
展开▼