| The problems caused by the conflict between the increasement of population and the limitation of resources promote the application of remote sensing (RS) and geographic information system (GIS) in agriculture, and precision agriculture (PA) is becoming the most important way to solve the problem of resource degradation and environmental deterioration and improve agricultural production efficiency. The development of PC desiderates exact information gained by RS rapidly (Moran et al., 1997), hyperspectral technology which acts as the hot and frontier of RS in the world, is showing outstanding potential in agriculture because of its ability not only to monitor crop and some of its surroundings timely, nondestructively in a big area as the normal RS technology does, but also improve the classification precision and accuracy of agriculture resources and supervise and analyses crop growth and the factors related with crop yield by spectral precision predominance. Soil is a foundation material of agricultural activity, and the data of some factors related with production are assurances of PC. Soil organic matter (SOM) is the most important constituent part of soil, can supply the nutrient element which plant needed, and also can improve soil construction and physical characteristics. The content of SOM in soil usually be a index of soil fertility, and it has become one of the focus to explain the law of reflectance changing with SOM and analyse SOM quantitatively by using hyperspectral. But, the hyperspectraapplication was limited because of the disunity of data collection and the weak of spectral energy and the interference of circumstance, And the existing researches have not yet concerned the influences on SOM hyperspectral characteristics caused by differences of minerals and some other materials of soil.The difference of the law that hyperspectral changes with SOM is the focus of the reseach. The influence on hyperspectral quality and repeatability by geometric conditions, the ways to manage sample surface and particle size was investigated in the study, and artificial SOM samples were gained to search the difference of hyperspectral characteristics, and the difference of the SOM prediction models of paddy soil derived different parent materials wasalso researched.The main contents of the study are as following:(1) Under the conditions of the same pretreatment for rice soil with close physical andchemical propertise,50W halogen lamp and 8 VOF, it indicates that the light source distance and angle significantly influence on the fluctuation and dispersion of hyperspectral data of soil by analyzing the influence of light source distance and angle, sensor distance to fluctuation and dispersion of hyperspectral data of soil indoor. There are differences in markedness of influcence at different waveband. The influence of sensor distance is unconspicuous. There are better hyperspectral data of soil indoor under testing geometric condition of 15?light source angle, 30 cm light source adistance and 15 cm sensor distance.(2) The surface process of soil sample and diameter of soil granule will obviously influence the dispersion of hyperspectral data of soil at four testing directions and average reflectance of five time. The influence degree is different for soil with different texture. The surface planishing and 1 mm diameter (where the soil sample passing through 1 mm standard griddle) are idealer process and appropriate diameter selection for measurement indoor, because they have lesser standard error vector of testing direction and average reflectance in repeated measurements for soil with different texture.(3)Keeping the same in three tache is the precondition for obtaining hyperspectral data ofsoil with higher generality indoor from gathering soil sample outdoor to measuring hyperspectral data indoor. Firstly, the soil samples are gathered from the same section such as surface layer soil samples gathered from 0 cm to 15 cm depth and representation on studied object. Secondly, the precondition of test...
|
PDF Full Text Request