| Agriculture is the foundation of national economy, and agricultural resources are the material bases and fundamental guarantee for the existence and development of agriculture that is very important to the human society. Prompt acquisition of the quantity, quality and changes of agricultural resource information is beneficial to the local conditions, reasonable planning and effective use of agricultural resources, and it is also quite helpful for sustainable development of agriculture, steady growth of rural economy, gradual income growth of farmers and obvious improvement of ecological environment. Taking Dingzhuang Town, Guangrao County of Shandong Province as a case, technology system about extraction of agricultural resource information was explored, which mainly including the soil physical and chemical properties, water resources, land resources, the planting area, growth condition and yield of crops, based on geographic information system (GIS), remote sensing (RS) and global positioning system (GPS). And then the spatial-temporal dynamic changes of agricultural resources were analyzed.(1) Extraction of soil physical and chemical properties'informationThe information of soil physical and chemical properties was acquired through inverse distance weighted (IDW) spatial interpolation of GIS based on the data of sampling points by GPS, which is the basis of calculating the precise requirement of N, P2O5 and K2O according to the nutrient balance model. The results showed that the spatial interpolation analysis was an efficient way to acquire the information of soil physical and chemical properties.(2) Extraction of water resource informationNir-Red spectrum feature space was established through image pre-processing, including relative radiometric calibration, atmosphere correction on TM Multi-spectral image of March 16,2009, and then information of water in the surface soil was extracted through model of SMMRS. The situation of ground water depth and mineralization was got by spatial interpolation analysis based on the data of sampling points. Surface water information was produced through TC transformation on remote sensing image.The results showed that the content of water in the surface soil of Dingzhuang Town is high, the average of which normalized reached to 0.87, and its spatial distribution accorded with the overall situation. Depth of ground water was low, basically less than 2 meters. Ground water mineralization was mainly 2～10 g/L, and even more than 30 g/L in costal area. Surface water density index was 9.52%.(3) Extraction of land resource informationChoosing the HJ1B Multi-spectral image of July 1,2009 which had significant features comparison, the land use types were obtained by decision tree classification based on the spectral characteristics of typical features, combining with expert knowledge and GIS, and then its results precision and land use degree were analyzed. Selecting four multi-spectral images in spring which have obvious saline-alkaline information, index model of lands saline-alkaline was established with the help of expert knowledge and GIS. Then the land saline-alkaline information was reached by the model, and its spatio-temporal dynamic changes were also analyzed. Basing on the analysis of environmental conditions of cultivated land, soil physical and chemical properties and agricultural infrastructure, indexes system was established with sufficient consideration of farmers'influence on cultivated land quality which was estimated by fuzzy discrimination method in 1979,2005 and 2007, and then their spatial and temporal variation characteristics were analyzed.The results showed that the precision for land use classification was 88.03%, and the high land use degree was 299.95.The area of land salinization was 62.60%,61.56%,59.51% and 71.53% of the total area in 1992,2001,2007 and 2009, respectively, and land salinization is common, most of which is mild and the area of the third and fourth grades salinization land has a large proportion. The average index of land salinization in the study area was 0.329,0.352,0.386 and 0.402 along the time, which had a tendency to aggravate as well as the districts. The land salinization degree declined gradually from east—north—south. The closer the lands are away from the Bo Sea, the lighter the land salinization was.Cultivated land quality of Dingzhuang Town improves continuously and presents obvious spatial-temporal variation rules. The area of the first and second grades is 0.060% of total cultivated land area in 1979, which increases to 91.42% and 98.62% in 2005 and 2007, separately, and the quality of cultivated land improves steadily. The quality of cultivated land decreases which locates far from administration center and rivers or nears the Bo Sea, and the increase amplitude of cultivated land that near the Bo Sea in the northeast and around the administration center is higher than other area. Average yield of crops have extremely significant linear relationship with cultivated land quality.(4) Extraction of crops'planting area, growth condition and the yield informationSelecting seven high and moderate resolution multi-spectral images, which have the similar temporal characteristics of obvious winter wheat information, the planting area of winter wheat was acquired by the decision tree classification approach through image pre-processing, including geometric correction, masking and relative radiometric calibration, based on the analysis of spectral characteristics of typical features in the study area. Vegetation indexes were calibrated according to their changing trends, and winter wheat yield estimation models were established based on∑RVI and DVI respectively according to their relationships with winter wheat yields.The results showed that high precision for winter wheat area extraction of more than 96%; precision for annual yield estimation was basically better than 90% with some fluctuation, and multi-year yield estimation based on∑RVI could reach 99%; precision for annual yield estimation and multi-year yield estimation based on DVI was more than 93.41%, and 97.01% respectively.This study provided a practical and effective technical system for agricultural resource information extraction especially at small scale, and it has great significance to the effective use of agricultural resources as well as the healthy and orderly development of agriculture and social economy.
|
PDF Full Text Request