Total Phosphorus Content In Panjin Paddy Field Based On Starland Remote Sensing

Rice is an important grain crop in China.Panjin City is a big rice planting city with1.4057 million mu of"paddy soil",accounting for 37.2%of the total soil area.Phosphorus is an essential nutrient for rice growth and one of the necessary elements to form protoplasts of plant cells.Phosphorus is involved in rice photosynthesis and the synthesis of sugar and starch in rice,which plays a decisive role in cell proliferation and growth.Therefore,accurate,efficient and rapid calculation of soil total phosphorus content and wide monitoring of the distribution of total phosphorus content in soil are the necessary prerequisites for the sustainable development of agricultural rice cultivation.In this paper,the total phosphorus content of paddy soil in the experimental area of Panjin saline-alkali land utilization research institute is studied.Through the interpretation of environmental satellite data and the analysis of field near-surface measured hyperspectral data,the spectral characteristics of total phosphorus content reflectance in paddy soil in Panjin rice field are revealed.The findings are as follows:(1)Compared with the original spectral data,the spectral differential transformation data have great differences,and the characteristic information is obvious.The general trend of spectral curves corresponding to different total phosphorus content is basically the same,showing an increasing trend.A peak appears at about 560 nm.A trough appeared around 690nm.Correlation coefficient between the original spectrum and soil total phosphorus content|r|≤0.35,and the maximum correlation coefficient of the original spectrum was increased by0.21 after the first and second order differential transformation.The partial least square regression model established by the second order differential logarithmic transformation of hyperspectral reflectance has a good prediction effect.The optimal model R~2is 0.672 and the RMSE is 0.126.Fitting check curve R~2is 0.742.(2)Satellite wide band reflectivity data and hyperspectral data fit well,soil spectral index and total phosphorus content correlation is good,correlation coefficient|r|≥0.5,comparative analysis of stepwise multivariate linear regression estimation model and partial least square regression estimation model,It is concluded that the R~2of stepwise multivariate linear regression model is 0.468,model RMSE is 0.206,partial least square regression model R~2is 0.569,model RMSE is 0.146.The partial least square model is selected as the quantitative inversion model of total phosphorus in Panjin soil.(3)The distribution of total phosphorus content in the soil in the early tillering stage of rice is that the total phosphorus content in the western region of Panjin is low,the total phosphorus content in the central region is moderate,the total phosphorus content in the coastal area and along the river area is higher,and the total phosphorus content in the paddy field area is increasing gradually.