基于最优模型的荒地土壤有机质含量空间反演  

Spatial inversion of soil organic matter content in wasteland based on optimal model

作　　者：乔娟峰 熊黑钢[2] 王小平[1] 周倩倩 

机构地区：[1]新疆大学资源与环境科学学院/绿洲生态教育部重点实验室,新疆乌鲁木齐830046 [2]北京联合大学应用文理学院,北京100083

出　　处：《江苏农业学报》2018年第1期68-75,共8页Jiangsu Journal of Agricultural Sciences

基　　金：国家自然科学基金项目(41671198)

摘　　要：本研究采用Landsat OLI多光谱遥感影像数据,结合实测土壤有机质含量,利用原始影像反射率(A)、反射率一阶微分(A')、反射率二阶微分(A″)建立单波段和多波段回归模型,估算研究区土壤有机质含量,反演其空间格局。结果显示,经微分处理后的影像反射率,与土壤有机质含量相关系数增大。其中A'处理后的遥感影像反射率与土壤有机质含量的相关系数达到-0.850,比原始的提高了0.401,增强了有机质的光谱信息。多波段回归建模效果优于单波段建模。且A'的多波段回归模型预测精度最好,其建模集R^2为0.80,RMSE为3.66,预测集R^2为0.79,RMSE为3.65,RPD为1.96,表明该模型精度高,误差最小,预测效果最优,可以很好地估算该区域的土壤有机质含量。基于一阶微分的多波段回归模型:SOM=23.12-470.94B3-24.35B4-43.06B6,对研究区的SOM含量空间分布格局进行反演,发现反演结果与实际情况吻合,因此,利用多波段回归模型能很好反演研究区SOM含量空间分布格局,表达其不同有机质含量的土壤空间分布与其对应的空间位置,这为土壤有机质面状参数的获取提供了快速而有效的方法。In this study,using Landsat OLI multispectral remote sensing image data,combined with the measured soil organic matter content,the original image reflectance( A),reflectance first order differential( A'),reflectance second order differential( A″) were used to establish single-band regression model to estimate the soil organic matter content and to retrieve the spatial pattern. The results showed that the correlation coefficient of image reflectance and soil organic matter content after differential treatment was increased. The correlation coefficient between the reflectance of the remote sensing image treated by A' and the content of soil organic matter reached-0.850,which was increased by 0.401 compared with the original,and the spectral information of organic matter was enhanced. Multi-band regression modeling was better than single-band modeling.And the prediction accuracy of the multi-band regression model was the best,and the modeling set R^2 was 0. 8,the RMSE was3. 66,the prediction set R^2 was 0. 79,the RMSE was 3. 65 and the RPD was 1. 96,which indicated that the model had optimal prediction effect. Based on the first-order differential multi-band regression model: SOM = 23. 12-470. 94 B3-24. 35 B4-43. 06 B6,the spatial distribution pattern of SOM content in the study area was retrieved and the result was consistent with the actual situation. The spatial distribution pattern of SOM content could be used to predict the spatial distribution of SOM content in the study area,and the spatial distribution of the soil with different organic matter content and its corresponding spatial location could be used to provide a quick and effective method for obtaining the surface parameters of soil organic matter.

关 键 词：影像反射率 土壤有机质含量 微分变换 多波段建模 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] S127[自动化与计算机技术—控制科学与工程]