Study On Positive And Opposite Degree Of Grey Incidence Indirect Estimation Model Of Soil Organic Matter Based On Hyper-spectral

Soil organic matter is an important component of soil,soil organic matter is and one of the important indexes to evaluate soil fertility.With the help of hyper-spectral technology to accurately and rapidly monitor the soil organic matter information of large-scale land,it has important guiding significance to promote the development of precision agriculture.The process of traditional soil organic matter measurement method is complex,low time efficiency,high cost.It is difficult to meet the needs of large-scale soil monitoring.Hyper-spectral remote sensing has become a new technology for soil organic matter monitoring due to its many bands,narrow bands and rich information.Due to the influence of many factors on soil spectral measurement,there is uncertainty in soil organic matter hyper-spectral estimation and the estimation accuracy is not high.Therefore,in this study,76 soil samples in surface layer and 76 soil samples in plough layer collected in Zhangqiu District,Jinan City,Shandong Province,were used as research objects.According to the uncertainty in soil spectral analysis,the research was carried out by means of cause analysis,statistical analysis and grey correlation analysis.The main research contents and conclusions are as follows:(1)The characteristic indexes of surface soil organic matter were determined.The spectral reflectance data after wavelet transform denoising and exception elimination were transformed by nine mathematical transformation methods.Following the principle of maximum correlation and making the feature bands as discrete as possible when selecting feature factors.The results show that the effects of transformation methods such as logarithm,reciprocal,and square root are not obvious.However,the first-order differentiation and its combination transformation have good results.The correlation between soil organic matter content and the transformed spectral data has greatly improved in some wavebands.Among them,absolute correlation coefficients of the logarithm reciprocal first-order differential method is more than 0.6 in the near 540nm,1200nm,1460nm-1650nm,1980nm-2130nm and 2280nm-2330nm,some of which are greater than 0.7,and the highest is 0.78.After the square root inverse first-order differential transformation,the correlation coefficient is higher at 800nm-880nm,and the highest reaches-0.70.557nm,1621nm,2107nm and2316nm were selected from the data transformed by the first order differential method of logarithm reciprocal.864nm was selected from the data transformed by the square root inverse first-order differential.Five bands were used as characteristic factors to estimate soil organic matter,and their correlation coefficients are 0.679,0.759,0.780,-0.694,and-0.700,respectively.(2)The positive and opposite degree of grey incidence estimation model of soil organic matter based on hyper-spectral data were establishedAccording to the difference of approaching direction between the estimated samples and the known patterns,this paper proposed the concept of positive and opposite degree of grey incidence and two novel models of that.The proposed model of soil organic matter based on hyper-spectral data was established.By comparing with the estimation results of the classical degree of grey incidence models and the traditional organic matter content hyper-spectral estimation methods,the mean relative error and determination coefficient R~2 of the area difference positive and opposite degree of grey incidence model are 5.312%and 0.930,and the index difference positive and opposite degree of grey incidence model are 5.911%and0.898.The accuracy is obviously superior to other comparison methods.(3)Indirect estimation of soil organic matter content in plough layer was realizedThrough the establishment of the scatter plot of the soil organic matter content in surface layer and plough layer and the correlation analysis of the organic matter content of in two kinds of soils,it shows that there is a significant correlation between the organic matter content in surface layer and plough layer.Then,according this correlation,the indirect estimation model of multiple functions of soil organic matter content was established,and the indirect estimation results of soil organic matter content in plough layer were obtained.The experimental results show that the accuracy of the indirect estimation of the soil organic matter in plough layer obtained by the area difference positive and opposite degree of grey incidence combined with the linear function relationship model is the best and the average relative error and determination coefficient R~2 are 9.782%and 0.819,respectively.