A Rearch About Remote Sensing Monitoring Method Of Soil Organic Matter Based On Imaging Spectrum Technology

In recent years,with the continuous development of hyperspectral technology,effective information acquisition target by spectrum technology has become the research hotspot in the current agricultural remote sensing.Soil organic matter content who is the important index of the arable land soil fertility,is one of the important material base to improve crop yields.Therefore,the quantitative study of the arable soil organic matter content by using the imaging hyperspectral remote sensing technology has become a hot spot in the application of remote sensing in agriculture.However,the research of the soil organic matter based on imaging hyperspectral technology is less in China.In view of this,the author designed the SOM Monitoring Experiment based on imaging hyperspectral.This paper takes the land soil in Yangling of Shaanxi Province,Anping of Hebei Province,Xiaotangshan of Beijing City as the research object,adopts spectral imaging technology,makes the spectral response law of arable soil organic matter content clear,introduces the resampling methods,soil index methods,principle components analysis(PCA),continuous wavelet transform(CWT),etc.,and carries out the remote sensing monitoring research of soil organic matter in small regional local scale.The research contents and the main research achievements of this paper are as follows:Based on soil index,principal component analysis and continuous wavelet transform,characteristic analysis of imaging hyperspectral data of soil organic matter content,carry out the response analysis of soil organic matter content and spectral parameters,extract the sensitive characteristics of different methods,determine the characteristic parameter of good correlation.According to the above feature analysis of the soil organic matter content,screen sensitive information of soil organic matter of different scales and different inversion algorithm,build predictive model which is based on soil index method,principal component analysis,continuous wavelet transform,and by the correlation coefficient,root mean square error and the estimation precision as evaluation index,verify the accuracy of the prediction model,and choose the optimal inversion model.By contrast,the accuracy of PCA-PLS model is the lowest,the accuracy of MSI-PLS model take second place,the highest accuracy is CWT-SLR model,and the inversion effect is best.The highest R2 value is 0.5762,the lowest RMSE value is 2.5547 g/Kg,and the overall estimate accuracy of EA is 87.19%.The uav flying height,pixel scale,imaging width has a certain contact.Different flying height,different pixel scale and different spatial resolution,lead to different precision of inversion model of arable soil organic matter.This paper simulated sampling data of five different pixel number,used to study the remote sensing model of arable soil organic matter.It constructs the relationship table of unmanned aerial vehicle(uav)flight level-imaging width-pixel resolution-inversion model-monitoring accuracy.See from the relationship table,the prediction model which the lowest flying height is 83 m is the best.Therefore,in the 5*5 sampling scale and the altitude of 83 m,this paper simulate the soil organic matter content with CWT-SLR model.In this study,based on the Imaging hyperspectral data,the spectral characters of soil organic matter were studied and the estimation models were established,and the rapid,accurate and non-destructive estimation technical process of the soil organic matter content was optimized and put forward.It improves the remote sensing monitoring system of soil organic matter in small regional local scale gradually,accelerate the development of precision agriculture and the applications of imaging hyperspectral technology which can provide decision-making guidance for that unmanned aerial vehicle(uav)imaging spectral technique applied to the soil organic matter.