| Cultivated land, is the essence of land resource.The quality of cultivated 1 is directly related to the human food security, social stability and long-term development.It has become an important subject to get the real information of cultivated land timely and evaluate and monitor the complex quality station and soil conditions Objectively using advanced technology and methods.Visible-Near Infrared (VisNIR) hyperspectral technology has a very high spectral resolution.Based on the analysis of the mechanism of soil physical and chemical properties and spectral response, it has the capability of quantitative for soil parameters. It provides the possibility for getting information of the cultivated land soil components effectively and accurately.In this paper, first of all, from the perspective of the farmland soil fertility index in the quality, considering the lack of soil fertility we introduced three indexes Total Nitrogen, available P, available K into the existing cultivated land quality evaluation index,and corrected the cultivated land quality evaluation results.Revised farmland natural quality index and the plot of the actual output correlation coefficient R2 increased from 0.782 to 0.852, suggesting that after joining the soil fertility index of cultivated land quality evaluation system is more effective than the original farmland classification system of land parcel scale of cultivated land quality.Secondly, this paper introduced high visible light and near infrared spectral analysis technology, try from two angles of and reverse to solve fast accumulating farmland component information and integrated quality information and then quickly survey and evaluation for cultivated land quality.Forward angle refers to the use of hyperspectral technology to estimate soil machine quality, total nitrogen, effective phosphorus rapidly-available potassium, pH fertility index, using traditional farmland classification and multi-factor weighted summation, calculating the farmland natural quality index (AQIG), from the perspective of improving the efficiency of farmland information to realize rapid evaluation on the quality of cultivated land.Articles used in the process of least squares to build model and tried a variety of soil reflectance data handling.The results shows that spectral reflectance of soil organic matter by the continuum removal (CR), total nitrogen by MSC+S.G smooth+second order differential transformation, effective phosphorus rapidly-available potassium after a second order differential transformation, effect of pH value after first order differential transform modeling and forecasting effect is best.In view of soil spectral data and farmland natural quality index are soil characterization of multi-factor index more interaction.Reverse attempts to establish a relationship between spectral reflectance data and index of cultivated land quality, so that we can use soil reflectance data predict farmland natural quality index (AQIF) directly.The inversion model established by using partial least squares between fertility index measured value measure farmland natural quality index AQIx and spectral reflectance with second order differential has the best conduct and prediction effect.R2 is 0.94, the prediction set RPD reached 3.64, which can realize fast and exact estimation of farmland natural quality index.Farmland natural quality index is concentrated in the visible light characteristic of band 520-760nm and Near Infrared band of 800-930 nm and 1450nm and 1700-2360nm range, which has common part with the particular range of soil organic matter, total nitrogen, effective phosphorus, rapidly-available potassium and pH.The result suggests that spectral reflectance data can not only estimate the response component content in the soil, and also estimate the height of the natural quality index which reflects the quality of cultivated land.Finally, this paper compares and analyzes the positive reverse Angle measuring farmland natural quality index.We found that the error between farmland natural quality index using hyperspectral data to estimate the soil fertility index AQIG and natural and other index AQI from the original farmland classification within the plus or minus 400, and the vast majority level of arable land have no change.The result means our analyses is good enough to keep up with the original grading results of inheritance. The correlation coefficient between AQIF and plot the actual standard food output is 0.81, higher than the original cultivated land quality index, which means AQIF can be used for cultivated land quality evaluation, and the effect is good.AQIG phase relationship with standard grain yield is only 0.73, the lowest in the four this may be due to different fertility index in the spectral estimation will have certain error, especially the content of soil effective phosphorus to estimate the effect is not very ideal, when calculating the farmland natural quality index the cumulative error,may lead to the correlation coefficient is not high. |
PDF Full Text Request