Font Size: a A A

Estimation Technology Of Soil Total Nitrogen And Its Influence Mechanism Based On Hyperspectral Inversion

Posted on:2021-01-15Degree:MasterType:Thesis
Country:ChinaCandidate:X SongFull Text:PDF
GTID:2393330602473154Subject:Agricultural resource utilization
Abstract/Summary:PDF Full Text Request
For the sake of the environmental quality of agricultural ecosystems,precise fertilization based on soil nutrients is crucial.Soil reflectance and mathematical statistical model were used to identify the hyperspectral characteristics of soil total nitrogen content.Hyperspectral reflectance provides a basis for rapid and accurate estimation and a technical reference for the promotion of precision agriculture by constructing a quantitative analysis model and exploring the response mechanism.Taking the peach plantation in binhu district,Wuxi city,Jiangsu province,and the experimental station of intensive agricultural ecosystem in north China,Huantai County,Shandong province,as the study areas,representative samples with large geographical span were selected for total nitrogen content determination and spectral data collection.The spectral reflectance was pretreated by the first-order differential method and the sensitive wavelength was selected by the peak valley method of the correlation coefficient.The MSASI was established by combining the two sensitive wavelengths.The calculated data were used to construct the quantitative hyperspectral analysis of soil total nitrogen content by MLR,BPNN,and PLSR,respectively analyze the model.After wheat and corn straw returning to the field,there were many differences in soil organic matter and different response laws.According to these distinctions,the optimal models were constructed.To explore the response mechanism of soil total nitrogen hyperspectral,the functional groups,soil minerals and surface morphology of soil were compared and analyzed by FTIR,XRD,and SEM.The results are as follows:(1)The hyperspectral sensitive bands of total nitrogen content of yellow brown soil were selected as 420~444 nm and 480~537 nm in Binhu research area of Wuxi City,Jiangsu Province,and the optimal pretreatment method was the first derivative.The best inversion model was the multiple linear regression analysis based on MSASI(R~2=0.98,RMSE=0.04).(2)The best pretreatment method of total nitrogen content of calcareous tidal soil was the first derivative transformation after the square operation in the North China intensive agricultural ecosystem test station of Huantai County,Shandong Province.The characteristic wavelengths of the soil in the wheat harvest seasons were 602,720,775,926 and 2215 nm,the soil in the corn harvest seasons was 442,553,998,1082,1087,1285,1314,1969 and 2377 nm.The determination coefficients(R~2)of BPNN in model set and verification set reached 0.80 and0.90 respectively.(3)The spectral index of hyperspectral data in different study areas was calculated,and the optimal spectral index was MSASI.It was speculated that the spectral index was more suitable for hyperspectral inversion of soil total nitrogen content in the test area.(4)For the soil of winter wheat-summer maize rotation system,the wheat straw of last season in the soil of corn harvest period was in the stage of accelerated maturity,and the corn straw in the soil of wheat harvest period was completely decomposed.The differences in soil physical properties caused by straw maturity were reflected in the hyperspectral characteristics of the soil,and the accuracy of the model was further affected by the different reflection characteristics.Therefore,hyperspectral inversion technology could be used for the long-term location study of soil fertility diagnosis,but it was necessary to fully consider the types of plant species,especially the straw returned to the field plot,which could provide a reference for the improvement of precision agriculture management.The results showed that the model selection and model accuracy of soil nutrient hyperspectral inversion were greatly influenced by crop planting type,and the soil nutrient estimation based on crop planting type classification was more accurate.Hyperspectral inversion technology provides a reference for the rapid diagnosis of soil total nitrogen content.Compared with the traditional soil nutrient determination method,it has lower cost and higher test efficiency.Its fast and convenient operation advantages can provide a technical reference for the accurate management of soil nutrients.
Keywords/Search Tags:Spectral reflectance, Soil nutrients, Spectral parameters, Straw returning to the field, Precision agriculture
PDF Full Text Request
Related items