| This study is based on Fukang wasteland soil in Xinjiang,firstly,64 soil samples were collected in the field,the content of organic matter was measured in the room,the spatial variability was analyzed,and the optimal parameters of half variance function were used,and in the area was made by using the common kriging interpolation in ArcGIS land statistic analysis,In order to obtain the method of predicting the spatial variability of soil organic matter content.Secondly,through the United States ASD spectrometer in the field to collect soil spectral reflectance as a data source,the spectral data smoothing and transformation after the first,analysis of soil spectral characteristics.Secondly,the characteristics of spectral curves under different organic matter levels were analyzed.Thirdly,the multivariate stepwise regression model of soil spectral data and its transformation form and organic matter content was used.Finally,using the remote sensing image Landsat 8,the paper extracts the ENVI soft surface reflectance,extracts the soil surface reflectance by IDL,then models it with the organic matter content,applies the optimal model to the Landsat 8 image,and obtains the spatial pattern distribution map of the soil organic matter content in the research area,In order to provide technical support for real-time monitoring of spatial distribution information of large scale soil fertility status in the region.The findings are as follows:(1)The soil organic matter of Fukang wasteland in Xinjiang is more barren,and its variation range is 0.59~23.55g/kg,which is lack of level.From the average,the soil organic matter quantity is 9.61g/kg,the overall low,belongs to the very deficient level,indicated that the area soil is barren and the nutrient content is few.(2)Using the half variance function to establish the soil organic matter content four kinds of models,the optimal fitting parameter is the Gaussian model.At the same time,the soil organic matter content was analyzed in the trend surface,and it was found that the organic matter in the east and West was obviously changed,showing a slant straight line,revealing that the content was gradually decreased from west to east with first order effect.The content of soil organic matter in north-south directionis high and has the second order effect.The spatial distribution of organic matter is obtained by the ordinary kriging interpolation.In the study area,the SOM content of soil surface is lower,and the content of organic matter is in the center of the reservoir dam,and the trend is increasing downstream.The SOM in the marginal area near the reservoir(southeast direction)is lower,and the SOM in northeast,northwest and south west dstrict of the research area is relatively high.(3)By analyzing the characteristics of the spectral curve of soil samples,the curve of the study area is consistent with the wave pattern in arid region.The 8spectral reflectance curves of parallel reservoir dams are roughly parallel and similar in shape,and the rate of reflectivity growth varies from 900 m,1200m and 1500 m curves of reservoir to 1000~1300nm interval.Roughly on the reflectance of the reservoir from the dam farther away,the reflectivity appears to be decreasing trend.At the same time,there is a vapor absorption zone near the 1400 nm,1900nm and2200 nm bands.(4)The spectral reflectance of different organic matter levels was analyzed,and it was found that the spectral reflectance of 3 levels of organic matter in near infrared band was larger than that of visible light,and the difference between curves was slightly greater than that of visible light.The spectral reflectance is the highest in the lack of grade,the spectral reflectance is the second,the spectral reflectance is the lowest in the lack of the range,and the higher the soil organic content is,the lower the spectral reflectance of the soil is.(5)Reflectance first derivative(FDR)for spectral reflectance(REF),after 5transformations of the reflectivity continuum removal(CR),the reflectivity continuum to remove the first order differential(CR'),the reflectivity reciprocal logarithm(Lg(1/R)),the reflectivity countdown logarithm First order differential(Lg(1/R)),It is found that the spectral reflectance after mathematical transformation enlarges the spectral band difference,which is most obvious in the visible light band.At the same time,the correlation coefficient of the organic matter content and the spectral reflectance of the soil is inversely proportional to that of the contents,and the correlation coefficients are promoted after the transformation of the reflectance,whichshows that some small spectral absorption characteristics are enlarged after the mathematical transformation.(6)In the significant band modeling,the SMLR model of REF,FDR,CR,CR',Lg(1/R)' and Lg(1/R),the model 1/r of ref,CR and Lg(RPD)are all 1.0~1.4,which can only make general prediction of organic matter content.The RPD of FDR and CR'model is better than that of 1.8~2.0 model.Lg(1/R)' established model of RPD is greater than 2.0,the model predictive ability is excellent,can accurately predict organic matter content. |
PDF Full Text Request