Research On Winter Wheat Planting Area Extraction Method Based On GF-1 Image

Winter wheat,as one of the main food crops in China,ensures the planting area plays an important role.In the past,the planting area of crops in China was mainly obtained through field measurement and then reported by administrative units as a unit.This method not only consumes a lot of manpower and material resources,but also does not guarantee the objectivity of the data,and does not obtain the specific spatial and temporal distribution of the cultivated land area.The use of remote sensing technology for crop area extraction can greatly reduce the cost of measurement while ensuring the objectivity and accuracy of the data.This paper takes Xiangfu District of Kaifeng City,Henan Province as the research area,and uses GF-1/WFV domestic data as the main data source.Google Earth(hereinafter referred to as GE)0.3m historical image is used as auxiliary image selection sampling point,respectively for single scene image and multi-time series images,study how to extract the planting area of winter wheat quickly and efficiently,the main research contents and conclusions are as follows:1.Based on GF-1/WFV single phase image during the growth cycle of winter wheat,the method of extracting winter wheat planting area in xiangfu region was studied.GF-1/WFV 5 field images during the winter wheat growth cycle in xiangfu district in 2015 were selected,and GE images close to the same period were used as auxiliary images to select sampling points.Before using the maximum likelihood method to extract the cultivated area of winter wheat in the study area,the salt film of land use type map was firstly applied to remove the cultivated area to improve the precision.The experimental results showed that the overall classification accuracy of these 5 phases was greater than 95%,and the Kappa coefficient was greater than 0.96.Three images of winter wheat growth cycle in 2014,2015 and 2016 were selected.Using the same method,the experimental classification accuracy and Kappa coefficient were also high.Moreover,the extraction area in the three years showed a change trend consistent with that published in the statistical yearbook.As a result,the image quality and crop growth cycle under the limitation of objective conditions,such as using GE images as auxiliary choose sampling points,the use of maximum likelihood method to extract the few area before winter wheat planting area,using the land use type figure cured film unless the land area,not only convenient for computer batch processing efficiency,and the result accuracy is also will be helpful to the farm,so as to realize the interannual change monitoring.2.Based on GF-1/WFV multi-temporal images during the growth cycle of winter wheat,the method of extracting winter wheat planting area in xiangfu region was studied.GF-1/WFV images of 14 landscape multiple time series in winter wheat growth cycle from October 2015 to June 2016 were selected.GE image was used as the auxiliary image to select the sampling points,and the pickled film was used to remove the cultivated area,and the characteristic values of winter wheat that were different from other landscape features were statistically analyzed.The sustained growth of NDVI value at the early growth stage of winter wheat was taken as the critical threshold to determine the conditions,and the decision tree model was established to extract the planting area of winter wheat in xiangfu area.The overall classification accuracy of the experiment was 96.85%,and the Kappa coefficient was about 0.71,indicating a high degree of consistency.Therefore,it is feasible to use the decision tree model proposed in this paper to extract winter wheat planting area in xiangfu district.Based on GF-1/WFV domestic satellite image as the main data source,before supervision and classification,the pre-treated image was marinized according to land use type map proposed in this paper,and the decision tree model established in this paper was used to extract winter wheat planting area in xiangfu district of kaifeng city.The experiment shows that the use of land use type map to remove unless cultivated area,not only can effectively remove unless cultivated area surface features of winter wheat area extraction errors,but also facilitate the computer batch processing.Taking GE image as the auxiliary image to select sampling points can effectively save the measurement cost and further improve the precision of winter wheat planting area.The accuracy of the experimental results can provide scientific basis for the formulation of agricultural policies by relevant local departments.